LobeHub: baixar versions anteriors

This release was automatically published from PR #21078.
Changes
See PR description: https://github.com/lobehub/lobehub/pull/21078
Commit Message
🚀 LobeHub Release (20261018)
Release Date: October 18, 2026
Since v2.2.19: 419 merged PRs · 14 contributors
Agents now reach further across your devices, and the app opens faster. Topics, documents and device data load from local copies first, and server runs can use models that only exist on your device.

✨ Highlights
- Goals with batch rollout — A goal stays a draft until you start it, and a Goal can roll out one repeated task in batches. Each batch stops at a decision gate for the owner to review. (#20840, #20704, #20917, #20833)
- Widgets & dashboards — Home and project dashboards, an agent widget tool, chat previews, and new lh widget / lh dashboard commands. (#20421, #20771)
- LLM relay to your device — Server-side runs, preset tasks, model lists and Ollama downloads can now use models that are only available in your browser tab or desktop app. (#20453, #20454, #20455, #20456, #20457)
- Remote terminals & device maintenance — Open interactive shells on connected devices from the web, and update or restart CLI daemons from device settings. (#20600, #20527, #20528)
- Local-first loading — Topics, agent documents, devices, git status and home recents load from local replicas first, so pages show content before the network responds. (#20449, #20869, #20838, #20841, #20851)
- Agent document folders — Folders, batch moves, documents as parents, and sorting by name, created or modified time. (#20761, #20926)
- Memory from your materials — Memory can extract rules from pasted text, files, documents and conversations. (#20266)
- Fork from a reply — Start a new topic from any assistant message. (#20610)

🏗️ Core Agent & Architecture
- Sub-agents run through the parent's executor, and afterToolCall controls can gate tool results. (#20446, #20297)
- Skills selected in the composer or CLI are injected on server runs. Adding an agent installs the Market skills it references. (#20961, #21059)
- Agent-to-agent messages now show which agent sent them. (#20548)
- Fewer repeated device and Market lookups on each send, and project sends appear right away. (#20465, #20897, #20628)
📱 Platforms / Integrations
- Heterogeneous agents: DeepSeek Harness adapter. Codex app-server can run on connected devices (behind Labs). Codex user messages can be edited and resent. Codex and Pi can use any tool-capable LobeHub model. Droid and OpenCode support server-default models. (#18333, #20820, #20467, #20612, #20965, #20963)
- Desktop: Bundled Computer Use app on macOS, guided macOS privacy permissions, Dock menu with running topics, tab drag with pin/unpin, and instant topic switching. (#20780, #20932, #20635, #20650, #21001)
- Models: Claude Haiku 5.5, fal H3 Max and Seedance 2.5 video. (#20517, #20210)
- OpenAPI: Async IM channel and push-token REST resources. (#20440)
🖥️ CLI & User Experience
- Tasks: archiving, creating tasks from Kanban columns, and creating or searching tasks from the command menu. (#20990, #21020, #21000, #21002)
- Sidebar topics can be grouped by device, time, status or agent within a project. (#20811, #20592, #20715)
- Projects can have a logo, and the trash can be filtered by project. (#20641, #20739)
- lh connect --auv offers Computer Use, and lh connect service install accepts --workspace. (#20983, #20818)
- Settings search ranks results better and handles IME input. (#20620)
🔒 Security & Reliability
- Security: Desktop can require security updates and checks the first launch of Canary builds. Credentials are collected through a secure form. nodemailer is updated to ^9.1.1. (#20168, #20469, #20338)
- Reliability: OAuth no longer goes through email validation, and replicas ignore hydrates and errors that arrive late. (#20545, #20754)

👥 Contributors
Huge thanks to 14 contributors who shipped 419 merged PRs this cycle.
@mise42 · @0genlab · @arvinxx · @tjx666 · @Rdmclin2 · @Innei 

This release was automatically published from PR #19137.
Changes
See PR description: https://github.com/lobehub/lobehub/pull/19137
Commit Message
🚀 LobeHub Release (v2.2.16)
Release Date: September 3, 2026
Since v2.2.15: 246 merged PRs · 9 contributors
This release makes long-running Goals easier to plan, supervise, and accept; adds direct Agent sharing and a project-scoped Acceptance workspace; broadens local heterogeneous Agent support; and strengthens search, CLI, collaboration, performance, and gateway reliability.

✨ Highlights
- More heterogeneous Agents, with safer bindings — Factory Droid joins through ACP, TRAE and Grok Build gain provider binding support, and the deployment-provided LobeHub connection expands across Claude Code, Codex, Kimi Code, Pi, and Grok Build. (#18737, #18775, #18785, #18811, #18932)
- Full-text search gains an Elasticsearch path — A provider-neutral search layer, durable outbox, resumable indexing, operational recovery, and an opt-in Docker Compose profile make Elasticsearch practical without changing the default self-hosted setup. (#18807, #18902, #19028)
- The CLI covers more of the Agent workflow — Workspace selection, resilient Goal execution, operation inspection and replay, execution snapshots, and stronger daemon readiness checks make lh more useful for repeatable local and remote workflows. (#18797, #18874, #19000, #19015, #19020)
- Workspace collaboration becomes more actionable — Document comment threads gain participant notifications and deep links, shared documents can receive likes, and Tasks add assignments, mentions, and member-aware tool actions. (#18750, #18999, #19035, #19039)
- Large workloads feel lighter — A CI-enforced first-screen graph cut, off-main-thread hashing, streaming upload paths, and larger PDF reads reduce startup and file-processing stalls. (#18945, #19033, #19065, #19073)
- Gateway and runtime failures recover more clearly — Device handshake deadlines, explicit device-read failures, recoverable offline-device states, bounded tool-loop handling, and synthetic tool-failure diagnosis prevent silent fallbacks and stuck execution. (#18965, #19021, #19022, #19069, #19116)

🏗️ Core Agent & Architecture
Goals, Tasks & Acceptance
- Goal planning now records exploration structure, schedules independent Tasks in parallel, and bounds Goal reads, lifecycle transitions, and deadlines. (#18942, #18944, #18970)
- Goal creation and execution stay non-blocking, while the Goal page adds contextual chat, liveness heartbeats, explicit process controls, structured acceptance criteria, and a final result-review view. (#18979, #19050, #19124)
- Acceptance lists support pagination and batch actions, and reports, decisions, and evidence are clearer during review and reruns. (#18843, #18844, #18896, #18897, #18901)
- Projects now host their own Acceptance workspace, including scoped queries, contextual empty states, and a notice when a shared viewer cannot perform review actions. (#19042, #19064)
Runtime, Context & Agent Work
- Agent Work verification no longer redispatches an already delivered Work item, and gateway tool-state updates are coalesced to reduce redundant traffic. (#18969, #19060)
- Repeated tool loops are bounded, synthetic tool failures retain their reason and tool identity, and network-success responses with empty completions can be retried safely. (#18965, #19083, #19116)
- Heterogeneous Agent lifecycle hooks persist across operations, while operation completion propagates through the hook chain. (#18408, #18946)
- Agent documents keep their folder relationships, and workspace resource management is less restrictive where ownership and permissions already allow the operation. (#18964, #19107)

📱 Platforms & Integrations
Heterogeneous Agents
- Factory Droid is available as a first-class ACP Agent with session resume, image input, cancellation, model configuration, and permission interventions. (#18775)
- TRAE supports provider binding through OpenAI Responses, incl

This release was automatically published from PR #18747.
Changes
See PR description: https://github.com/lobehub/lobehub/pull/18747
Commit Message
🚀 LobeHub Release (20260827)
Release Date: August 27, 2026
Since v2.2.14: 235 merged PRs · 13 contributors
This cycle makes long-running Agent work easier to direct and recover, expands provider-backed heterogeneous Agents, delivers safer restart-free Desktop updates, and deepens Artifact, Workspace, and acceptance workflows.

✨ Highlights
- Provider-backed heterogeneous Agents — Claude Code and Codex gain deployment-backed model options, while Kimi Code and Pi add new provider paths across web and Desktop. (#18468, #18558, #18609, #18658, #18720, #18721)
- Restart-free Desktop updates — Signed renderer updates apply in place with compatibility checks, compressed delta delivery, release-scoped artifacts, and automatic rollback. (#18648, #18694, #18716, #18724)
- Agent approval and intervention flows — Agent approvals now align across surfaces, and generic interventions persist so long-running work can resume with the right user decision. (#18417, #18645)
- Agent self-learning — A Labs-gated growth portrait shows learned domains, rules, and practice history while operation-scoped expertise improves future Work. (#18195, #18394, #18420)
- Public Artifacts and Workbench — Local HTML and site assets can be packaged into versioned Artifacts, while acceptance and verification run in a standalone SSR Workbench. (#18424, #18473, #18569, #18728)
- Workspace collaboration — Per-member Resource grants, transfer requests, and a redesigned Notification Center make shared work easier to control and act on. (#18389, #18432, #18536, #18545)
- More adaptable everyday UI — Global interface fonts, consolidated Home promotions, clearer model availability, and refined recommendation visuals make the app easier to personalize and navigate. (#18712, #18722, #18730, #18741)

🏗️ Core Agent & Architecture
Goals, Tasks & Acceptance
- Goal state moves to a durable graph with dependency-aware scheduling, CLI controls, autonomous recovery, and terminal Acceptance Work. (#18574, #18597, #18635)
- Scheduled Tasks get a dedicated view, successful subtasks complete automatically, and Kanban grouping matches the list view with clearer loading states. (#18586, #18591, #18631, #18707)
- Acceptance deliveries group by Project and can be merged without losing checks, evidence, or prior review decisions; the review round now always reports its outcome. (#18448, #18537, #18548, #18672)
- Acceptance features now share one consistent product boundary, while abandoned Goal Work and stale scoped tool calls recover instead of silently ending the run. (#18656, #18664, #18703)
Agent Learning & Identity
- Self-learning gains dedicated domain creation, editable anchors, operation-scoped Context, and safer lesson history management. (#18418, #18420, #18517, #18556)
- Agent profiles gain a character sheet and reusable full-body artwork that retains its style and direction. (#18446, #18464)
- Agent authors can restrict Topic sharing, while Agent transfers preserve accumulated expertise. (#18445, #18557)
- Approval workflows align with web behavior, and generic interventions persist across Agent runs. (#18417, #18645)

📱 Platforms & Integrations
Heterogeneous Agents
- Claude Code supports local Provider bindings, while supported Claude Code and Codex runs can use deployment-backed models with operation-scoped authorization. (#18468, #18558, #18595, #18609, #18621)
- Kimi Code joins heterogeneous Agent provider bindings, and Pi gains a dedicated Desktop provider path. (#18720, #18721)
- Each Topic keeps its own heterogeneous model choice instead of rewriting the Agent default. (#18658)
- Cursor moves to a bidirectional ACP session with permission prompts, questions, approvals, and safe resume recovery. (#18475)
- OpenClaw and Hermes detection now covers official install paths, refreshed shell environments, and Windows launchers. (#186

This release was automatically published from PR #18373.
Changes
See PR description: https://github.com/lobehub/lobehub/pull/18373
Commit Message
🚀 LobeHub Release (20260816)
Release Date: August 16, 2026
Since v2.2.13: 363 merged PRs · 21 contributors
This cycle gives longer-horizon work a durable home: Projects and Goals land as first-class containers for agent work, seven new CLI coding agents join the heterogeneous runtime, a Local Sandbox executes on your machine, and the platform opens up through a generated OpenAPI spec and SDK.

✨ Highlights
- Seven new CLI coding agents — Cursor, CodeBuddy, Qoder, Kimi Code, Pi, TRAE, and Grok Build (ACP) join the heterogeneous runtime, plus a Codex app-server lab. (#18229, #18219, #17965, #18228, #17899, #18292, #18254, #18275)
- Local Sandbox — A local execution environment for devices, with a real working directory instead of a refusal. (#18143, #18180)
- Public API & SDK — openapi.yml generated from routes, a new @lobehub/sdk package, expanded v1 endpoints, and scoped API keys. (#17944, #17980, #18141, #18029)
- Home customization — A Customize modal with Minimal / Balanced / Full presets, a day-paged daily brief, and a task-shaped task mode with scheduled tasks. (#17912, #17948, #17958, #18173)
- Desktop split views — Split tab views, a tab strip reworked for many-tab use, and a faster cold start. (#18004, #17809, #17811)

🏗️ Core Agent & Architecture
Projects & Goals（Alpha）
- Project schema, backend and CLI, workspaces with conversation PoC, and an identifier column. (#17996, #18006, #18020, #18034)
- Goal creation split from task creation with an explanatory empty state; goals get their own table and AI-generated acceptance criteria. (#18047, #18335, #18265)
- Goal launch requires confirmation even in auto-run mode. (#18111)
Agents & Groups
- Group member permissions with personal model choice on shared builtins. (#18122)
- Async agent transfer and copy, with history backfill and intact membership lifecycle. (#17997, #18030, #18082, #18126)
- Sub-agents follow the parent model by default, with thinking controls. (#17938)
- Agent quota gets a usage calendar and burnout projection. (#18117)
Context & Memory
- Graph runtime context is injected continuously; container message tokens count toward the context budget. (#17761, #17839)
- Every memory tool call renders a dedicated card; extraction prompts align with their schemas. (#17891, #17988)
- Opt-in automatic topic summary workflow. (#17796, #18171)

📱 Platforms & Integrations
Heterogeneous CLI Agents
- New runtimes: Cursor CLI with model selection, CodeBuddy, Qoder with reasoning effort, Kimi Code, Pi, TRAE, and Grok Build over ACP; Codex app-server session lifecycle completes as a lab. (#18229, #18244, #18219, #17965, #18160, #18228, #17899, #18292, #18254, #17907, #18275)
- The model selector is driven by a capability table, with local CLI descriptors centralized as a single source of truth. (#18214, #17979)
- Windows support hardened: CLI spawning, agents behind unknown shims and stale PATH. (#18095, #18220)
- A heterogeneous error taxonomy classifies terminal errors; cloud auth failures surface clearly. (#17887, #18040)
Desktop
- Split tab views, a tab strip reworked for many-tab use, and pinned tabs that travel. (#18004, #17809, #17857)
- Boot straight into the last active workspace; cold-start critical path reduced and the local database prewarmed after navigation. (#17946, #17811, #18218)
- Chromium zoom presets and update downgrades. (#17679, #18012)
Messaging & Bots
- /mode command switches bot conversation mode; recent same-channel history is injected. (#18197, #16608)
- Proactive messenger push with per-channel notification settings. (#17791)
- Discord completes deferred forwarded interactions. (#18169)
Models
- Gemini 3.7 Flash, GLM-5.3 with always-on thinking, Grok 4.6 reasoning effort, and MiniMax-H3 video with official v2 API. (#18289, #18290, #18301, #18225, #17827)
- Audio lands in multimodal understanding wi

This release was automatically published from PR #17807.
Changes
See PR description: https://github.com/lobehub/lobehub/pull/17807
Commit Message
🚀 LobeHub Release (20260801)
Release Date: August 1, 2026
Since v2.2.12: 86 merged PRs · 13 contributors
Home becomes an Agent dashboard, the desktop shell keeps every tab alive, and local Agents connect through one wizard — with a long tail of runtime, model, and Windows fixes behind it.

✨ Highlights
- Home as an Agent dashboard — Home is rebuilt around what your Agents are doing: activity, Topics, and a dashboard rail replace the old landing view. (#17680, #17756, #17798, #17735, #17781)
- One wizard for local Agents — Connecting a local Agent now runs through a single unified flow instead of per-connector setup paths. (#17794)
- Local MCP connectors on desktop — Install and run stdio and LAN-HTTP MCP connectors directly from the desktop app. (#17673)
- Per-tab keep-alive routers on desktop — The Electron shell routes each tab through its own keep-alive router, so switching tabs no longer resets in-progress views. (#17595, #17773)
- Cross-runtime local database — Local data now runs on one database layer shared across runtimes, replacing the runtime-specific paths. (#17682)
- Scoped tool allowlists in the Agent runtime — Tool access is enforced against the calling scope, so a sub-Agent can no longer reach tools outside its grant. (#17717)
- Faster Agent search — Agent filtering moved above the BM25 scan and scans are isolated per query, restoring real TopN ordering and true relevance scores. (#17755, #17720)
- Demand-aware SPA startup — The SPA loads work on demand instead of up front, cutting cold-start cost. (#17577)

🏗️ Core Agent & Runtime
Conversation correctness
- Sub-Agent final answers are preserved when steps are grouped. (#17725)
- Explicit model overrides survive in sub-Agent Topics. (#17766)
- Assistant step count now matches the actual number of LLM calls. (#17790)
- Assistant messages are reused on step retry instead of duplicated. (#17703)
- Duplicate assistant continuations are omitted. (#17726)
- Failed assistant placeholders are filtered, and the Claude prefill guard is hardened. (#17737)
- Thread conversation cancellation is isolated, so cancelling one thread leaves the others running. (#17719)
- Runtime todos are injected into the server Context. (#17671)
Data & workflows
- Local Agent Signal workflows are supported. (#17681)
- Agent documents accept backing IDs when reading. (#17763)
- GitHub Works are registered from heterogeneous shell gh runs. (#17666)

🖥️ Workspace & Home
- Workspace budget settings get their own tab and route slot. (#17775)
- Onboarding generates starter task recommendations, with the recommendation prompts centralized. (#17757, #17770)
- Agent Profile supports a Markdown source mode. (#17580)
- Messenger DM support, plus broader Workspace interaction polish. (#17771)
- Claude Code quota is sampled from bound devices and refreshed every two minutes. (#17797)
- Settings page headers are unified, the About layout is reworked, and Workspace and service-model pages are tidied. (#17667, #17696, #17698)
- Memory empty states, the prompt editor header, and Agent statistics are polished. (#17764)
- Conversation portal headers are refined. (#17751)
- Model Select surfaces stale-model state instead of a raw provider/model id, and the stale-model popup is compact. (#17748, #17754)
- Tab titles show the section and Agent name on Agent sub-pages. (#17711)
- The Home sidebar stays put on routes without a dedicated panel. (#17718)
- Sidebar task list cache persists, so the skeleton no longer replays on every visit. (#17706)
- The Inbox route stays valid during initialization. (#17759)
- Resolved briefs no longer reappear after navigation. (#17760)
- The prompt view-mode toggle no longer scrolls to the bottom. (#17746)
- Enter / Esc hotkeys are added to the AskUserQuestion intervention. (#17733)
- A copy-operation-id action is available in assistant m

This release was automatically published from PR #17665.
Changes
See PR description: https://github.com/lobehub/lobehub/pull/17665
Commit Message
🚀 LobeHub Release (20260728)
Release Date: July 28, 2026
Since v2.2.11: 92 merged PRs · 17 contributors
This weekly release makes Agent work easier to review and recover, with structured verification, broader model access, clearer completion feedback, and a smoother desktop experience.

✨ Highlights
- Acceptance review flow — Open focused checks, review structured metrics and charts, and accept or reject delivery from one unified surface. (#17566, #17631, #17633, #17650)
- Chat image generation — Let chat models discover image models, inspect parameters, start generation, and follow generation status through a built-in tool. (#15332)
- More model access — Use a ChatGPT subscription through device authorization, run Claude Opus 5, and select the latest Gemini Flash models. (#17527, #17582, #17608, #17433)
- Agent work visibility — Turn generated files into durable Works, summarize edited files, and notify users when long Agent runs finish. (#17423, #17331)
- Desktop file preview — Preview local HTML work with its relative assets while keeping file access inside the approved project root. (#17545)

🏗️ Agents, Tasks, and Verification
- Unified Task criteria, check results, focused review comments, and accept or reject actions across desktop and mobile. (#17631)
- Added versioned structured verification data for metric comparisons, curves, scatter plots, heatmaps, SOTA tables, and grouped bar charts while retaining the raw evidence. (#17633)
- Kept shorter Acceptance checklists flat, grouped only larger lists, and protected owner-only run history in shared views. (#17650)
- Kept scheduled runs visible in active conversations and triggered recall for success-like completion states, including auto-repair outcomes. (#17473, #17613)
- Preserved human-intervention message ancestry after server-runtime rehydration so submitted or skipped responses stay attached to the requesting assistant. (#17628)
- Synchronized Task editor state and grouped retries, and restored image paste in the browser runtime. (#17585, #17644)

🤖 Models and Generation
- Added ChatGPT subscription authentication through the Codex OAuth device flow and a provider-scoped Responses runtime. (#17527)
- Added Claude Opus 5 runtime and model-card support, including adaptive-thinking defaults and summarized thinking display. (#17582, #17608, #17620)
- Sanitized replayed reasoning for Anthropic-compatible providers so missing text and foreign signatures do not break cross-model conversations. (#17629)
- Added Gemini 3.6 Flash and Gemini 3.5 Flash-Lite, preserved Google image resolution when aspect ratio is automatic, and corrected range-priced GPT model costs. (#17433, #16677, #16991)
- Added a built-in chat image-generation tool that reuses the existing model discovery, generation, and status pipeline. (#15332)

🖥️ Desktop and User Experience
- Added Preview and Source modes for local HTML files, including project-scoped asset resolution for CSS, JavaScript, images, fonts, audio, and video. (#17545)
- Moved the Working Sidebar behind a shared feature boundary and filtered OS metadata, dependency directories, caches, generated artifacts, and editor backups from project files. (#17646)
- Rendered gitignored project files in the Agent file browser and retained desktop browser webviews in the renderer. (#17517, #17513)
- Unified development tools in a bottom DevDock and restored React DevTools through a standalone desktop bridge. (#17499, #17602)
- Preserved Windows PATH order when locating Claude Code and Codex, and simplified terminal controls and Topic context menus. (#17381, #17609, #17642)

🔒 Security and Reliability
- Security: Kept local HTML preview resources inside the approved project root after symlink resolution. (#17545)
- Reliability: Restored correct provider error handling for Ollama and woke registered-only gat

This release was automatically published from PR #17509.
Changes
See PR description: https://github.com/lobehub/lobehub/pull/17509
Commit Message
🚀 LobeHub Release (20260723)
Release Date: July 23, 2026
Since v2.2.10: 366 merged PRs · 17 contributors
This release makes Agent work easier to run, inspect, and recover across the desktop, CLI, Workspace, and connected tools.

✨ Highlights
- Desktop terminal and browser workspace — Run terminal sessions beside a conversation, tune terminal fonts, capture browser screenshots into chat, and give Agents richer page context. (#17197, #17455, #17442, #17163)
- Local coding session continuity — Import Claude Code and Codex sessions, resume reclaimed Claude Code sessions, and follow live sub-agent progress. (#17101, #17437, #17115)
- Workspace access controls — Manage workspace-scoped API keys, resource permissions, connector ownership, device sharing, and private Agent access. (#17263, #17192, #17286, #17037, #17417)
- Broader model and provider support — Adds Kimi K3, SuperGrok, Gemini 3.6 Flash, Gemini 3.5 Flash-Lite, and OpenCode Agent support. (#17267, #17001, #17440, #17345)
- Topic-level model choice — New topics snapshot their selected model and keep model changes scoped to that topic. (#15933)

🏗️ Agents, Chat, and Workspace
- Added account-scoped Agent quota data, capacity calibration, selected-account routing, and per-turn usage tracking. (#17314, #17441)
- Added configurable default models for sub-agents and included sub-agent spend in parent usage. (#16552, #17116)
- Added topic scheduling, resumable heterogeneous Agent continuations, and forwarding topics or messages to Agents. (#17127, #17077, #17039)
- Improved long-topic reliability with newest-first message paging and a Topic doctor that finds and repairs hidden messages. (#17313, #17145)
- Added Workspace resource collaboration, RBAC controls, data deletion, and safer cross-workspace transfer behavior. (#17172, #17263, #17180)
- Improved Agent recovery by preserving rate-limit state, surfacing process errors, and timing out stalled connected Skill calls. (#16974, #17248, #17454)

🖥️ Desktop, CLI, and Verification
- Graduated the built-in terminal from Labs and added tabs, context menus, WebGL rendering, and persistent terminal settings. (#17459, #17494, #17455)
- Added an opt-in sandbox runtime for local device execution with explicit capabilities and launch policy. (#17476)
- Added browser control for desktop Agents, topic-scoped browser sessions, and inline browser screenshots in Agent output. (#17057, #17200, #17129)
- Expanded lh verify and the acceptance workflow with installed testing Skills, complete pageable topic views, and direct verification URLs after evidence submission. (#17280, #17224, #17460)
- Added OpenCode and Amp CLI Agent support, plus more reliable Codex todo streaming and Claude Code session recovery. (#17345, #17223, #17329, #17437)

🔌 Models and Integrations
- Added Gmail and GitHub connector data clients and agent-scoped connectors with Workspace isolation. (#17322, #17319, #16971)
- Added WeChat system bot support and Telegram slash-command handling. (#17275, #16545)
- Added Kimi K3, SuperGrok through xAI OAuth, Gemini 3.6 Flash, Gemini 3.5 Flash-Lite, and Grok prompt-cache support. (#17267, #17001, #17440, #17489)
- Improved model safeguards for unavailable image-generation models, thinking-only Qwen models, and unsupported MiniMax image detail values. (#17410, #17363, #17281)

🔒 Security and Reliability
- Security: Prevented MCP stdio pre-checks from inheriting the server process environment. (#17323)
- Security: Added workspace-scoped API keys with RBAC and masked real API-key prefixes in OpenAPI authentication errors. (#17263, #17143)
- Reliability: Reduced mobile stack-overflow risk in conversation traversal and prevented mobile input auto-zoom. (#17457, #17013)
- Reliability: Fixed stale running-operation reconnect loops, empty completion retries, and AI model batch-update duplicati

🚀 LobeHub Release (20260710)
Release Date: July 11, 2026
Since v2.2.9: 380 merged PRs · 385 commits · 16 contributors
This release makes Agent work more capable and dependable across local, remote, and scheduled runs, with clearer verification, stronger Workspace controls, and faster everyday chat workflows.

✨ Highlights
- Latest model support — Adds GPT-5.6, Grok 4.5, Sonnet 5, and Nano Banana 2 Lite metadata, with Codex usage support across Agent and CLI flows. (#16957, #16966, #16992)
- Worktree workflows — Adds worktree selection and creation directly from chat, with safer branch checkout, switching, and removal controls. (#15898, #16810, #16908, #16977)
- Claude Code and Codex interoperability — Improves remote human-in-the-loop, tool rendering, session recovery, quota controls, and image-result handling for heterogeneous Agents. (#16557, #16625, #16786, #16855)
- Downloads hub — Adds a central place to find LobeHub apps, CLI, and Messenger downloads. (#16929)
- Chat productivity — Adds input history, message multi-select and forwarding, file drag-and-drop, text-selection actions, and restore-to-input for CLI Agent turns. (#16400, #16446, #16555, #16558, #16842)

🏗️ Agent Runtime & Automation
- Moves core tool, sub-Agent, compression, retry, and completion executors into the shared Agent runtime for more consistent behavior across server and CLI execution. (#16914, #16925, #16945, #16948)
- Stabilizes QStash-backed runs by improving step startup, lock retries, guard cancellation, terminal event handling, and scheduled-task recovery. (#16721, #16829, #16832, #16907, #16716)
- Adds richer task run cards, inline long-run output, per-run follow-up, and clearer error states when verification infrastructure fails. (#16751, #16819, #16696)
- Adds device-scoped filesystem Skills and improves offline-device fallback, browser lifecycle handling, and local binary discovery. (#16682, #16759, #16888, #16952)

🖥️ CLI, Desktop & Chat
- Adds interactive code-context selection, project file search, concurrent desktop development instances, and dedicated inspectors for common coding commands. (#16681, #16526, #16553, #16662)
- Keeps streaming model, usage, tool, git, and session metadata synchronized through heterogeneous Agent runs. (#16963, #16904, #16592)
- Improves topic and sidebar consistency by fixing optimistic creation, activity ordering, unread indicators, Workspace isolation, and stale running states. (#16467, #16868, #16961, #16411, #16816)
- Graduates the floating chat panel from Labs and improves task-topic chat, Agent Documents, and review-panel workflows. (#16978, #16692, #16578, #16695)

🔌 Models, Integrations & Workspace
- Separates Skills and Connectors in Settings, exposes the full Connector catalog, and makes connection state and Workspace scope clearer. (#16975, #16789, #16633, #16620)
- Adds Weixin business bot support, locale-aware notifications, Composio mobile routes, and configurable AIHubMix proxy routing. (#16803, #16867, #16779, #16418)
- Adds Grok 4.5 reasoning-effort controls and lists embedding models in the service selector. (#16921, #16857)
- Adds Workspace visibility schema, personal-to-Workspace credential sharing, and organization credential routing for Agent runs. (#16598, #16895, #16889)

⚡ Performance, Reliability & Security
- Reduces Agent and chat latency through faster QStash publishing, parallel cache hydration, batched heterogeneous-message writes, lighter home payloads, and unread-message prefetching. (#16982, #16951, #16873, #16781, #16831)
- Makes fetch failures actionable with shared async boundaries and explicit retry states instead of silent empty views. (#16639, #16640, #16647, #16659)
- Fixes security boundaries for SSRF-safe Skill imports, knowledge-base file access, group ownership checks, and dependency updates. (#16601, #16594, #16586, #16041)
- Hardens Unicode and base64 handling for browser utilities, including non-ASCII text and valid surrogate pairs. (#16879, #16880)

🗄️ D

This release was automatically published from PR #16406.
Changes
See PR description: https://github.com/lobehub/lobehub/pull/16406
Commit Message
🚀 LobeHub Release (20260629)
Release Date: June 29, 2026
Since v2.2.8: 199 merged PRs · 199 commits · 14 contributors
This weekly patch strengthens the agent execution loop: tasks can be verified with evidence, heterogeneous runs recover more reliably, and desktop, CLI, workspace, and provider workflows get clearer control surfaces.

✨ Highlights
- Heterogeneous agent reliability — Routes gateway and external-agent runs through shared lifecycle hooks, persists operation traces, records provider/model details, and keeps bounded verification armed. (#16102, #16193, #16237, #16302, #16386, #16393)
- Task and group continuity — Keeps creator context, assignee activity, group identity, real member rosters, and final answers stable across task handoffs and multi-agent conversations. (#16162, #16188, #16213, #16228, #16303, #16315, #16352, #16363, #16385, #16399)
- Desktop and CLI control — Adds CLI self-update and verify commands, desktop completion notifications, overloaded Claude Code auto-retry, tab/window fixes, hotkeys, and safer local search limits. (#16052, #16147, #16164, #16190, #16277, #16281, #16309, #16384, #16396)
- Provider and model compatibility — Updates OpenAI SDK/runtime mapping, Bedrock key auth, contextual pricing, GLM-5.2 handling, Zhipu stream controls, Copilot parameter filtering, Kimi thinking preservation, and Doubao Seed models. (#16090, #16128, #16135, #15722, #16204, #16240, #16284, #16331, #16401)
- Connectors and integrations — Improves custom connectors, legacy MCP migration, context-aware tool manifests, Aihubmix base URLs, Search1API auto engine selection, PostHog skill integration, and MCP marketplace handling. (#15912, #16172, #16180, #16210, #16233, #16311, #16328, #16282)

🏗️ Core Agent & Architecture
Task and Verify
- Task lifecycle — Task creation now records creator conversation origin, supports task-level verify storage, links created task IDs to detail pages, includes subtask activity, and bridges finished task handoff back to the creator conversation. (#16111, #16113, #16213, #16238, #16296, #16303, #16399)
- Verification reports — Verify runs gain standalone ingestion/init/delete, evidence submission, acceptance injection, final report pages, table/mobile report layouts, and builder-captured evidence. (#16132, #16140, #16163, #16164, #16187, #16214, #16260, #16391)
- Agent Builder context — Builder suggestions, feedback loops, document-to-skill conversion, editing-agent context, and heterogeneous-agent discovery make agent setup and repair more direct. (#15939, #16225, #16292, #16326)
Conversations, Groups, and Runtime
- Shared run lifecycle — Chat, gateway, and heterogeneous execution continue converging on the shared lifecycle with transport-driven completion, unified run modules, auto-title generation, and terminal error surfacing. (#16083, #16086, #16102, #16119, #16151, #16155)
- Group identity stability — Group reads, persisted messages, supervisor replies, council rendering, member roster injection, and topic routing now preserve group identity across nested and server-runtime flows. (#16149, #16152, #16154, #16188, #16228, #16259, #16271, #16313, #16352, #16363)
- Tool-heavy turn handling — Conversation grouping keeps toolless prose, terminal errors, trailing bookkeeping tools, and post-tool answers readable without hiding the final answer. (#16084, #16127, #16278, #16377, #16385)

📱 Platforms & Integrations
Desktop, CLI, and Devices
- Desktop UX — Desktop tabs, dynamic route meta, workspace-scoped navigation, completion notifications, update prompts, working directory actions, and transient network handling are more predictable. (#16089, #16184, #16190, #16206, #16224, #16252, #16273, #16276, #16309, #16355, #16384, #16396)
- CLI workflows — lh update, verify skill initialization, verify run deletion, file-search result caps, worker 

This release was automatically published from PR #16183.
Changes
See PR description: https://github.com/lobehub/lobehub/pull/16183
Commit Message
🚀 LobeHub Release (20260622)
Hotfix Scope: Task template recommendation regression — legacy cached or server rows crash the home screen
Drops stale or malformed task-template recommendations before the Daily Brief UI renders them, restoring the home screen for version-skewed clients.
🐛 What's Fixed
- Home crash from legacy task templates — Legacy or malformed task-template recommendation rows are now filtered before connector-aware rendering and cache mutation, so stale SWR data or older server payloads no longer crash the Daily Brief section. (#16170)
⚙️ Upgrade
- Self-hosted: pull the new image and restart. No schema or env changes.
- Cloud: applied automatically.
👥 Owner
@tjx666

This release was automatically published from PR #16080.
Changes
See PR description: https://github.com/lobehub/lobehub/pull/16080
Commit Message
🚀 LobeHub Release (20260620)
Release Date: June 20, 2026
Since v2.2.6: 88 merged PRs · 7 contributors · 1 database migration
This week pairs a workspace-scoped database migration with broad runtime and UX work — native ASR transcription, audio attachments for audio-capable models, a smarter input-completion prompt, and a long list of reliability fixes.

🗄️ Database Migration (Action for Self-Hosted)
This release ships migration 0111_workspace_device_and_ai_infra_surrogate_pk (#16065). It is the headline change for operators — read this before upgrading.
What changes
- ai_providers / ai_models — adopt a surrogate _id primary key and workspace-scoped partial unique indexes, so provider/model identity is unique per workspace rather than per user.
- devices — the old (user_id, device_id) unique is replaced by two workspace-scoped partial uniques, letting personal and workspace-enrolled devices live in independent identity spaces.
- workspaces — adds the freeze trio frozen (boolean, default false), frozen_reason (text), frozen_at (timestamptz) backing workspace freeze risk control.
Operator impact
- Backwards-compatible and additive — every statement is guarded (IF EXISTS / IF NOT EXISTS / catalog checks). It is a NO-OP on cloud production (where the ai_infra side was already applied online) and a full rebuild on fresh / self-hosted databases.
- No manual SQL required — the migration runs automatically on application startup.
- Schedule the rollout in a low-traffic window and take a backup snapshot before deploying.
- If the migration fails, do not retry repeatedly — inspect migration logs and lock state first.
Rollback
Migration is additive only (new columns / indexes / surrogate PK). If your environment requires strict reversibility, follow your standard DB restore or migration-rollback policy from the pre-deploy snapshot.

✨ Highlights
- Native ASR transcription — new ASR (transcribe) runtime with OpenAI + Gemini-native backends, exposed via tRPC and the CLI (incl. remote audio URLs). (#15992, #15987)
- Audio in chat & agent mode — pass audio (mp3) attachments to audio-capable models like Gemini, with upload allowed in agent mode and a waveform player on render. (#15988, #16022)
- Input completion v1.2 — reworked completion prompt that predicts intent, including long-range continuations. (#15961)
- Market recommendations — task-template recommendations from Market, plus skill recommendations during agent creation. (#15558, #16016)
- Model intelligence — inject model knowledge cutoff into context and support GLM-5.2 reasoning effort. (#16048, #15972)
- Topic lifecycle — persist topic unread as a backend status and show a [Draft] hint on topics holding unsent input. (#16056, #15996)
- Workspace-scoped devices — distinguish local vs. gateway device targets, add a recent-directory button, and auto-activate a device for bot triggers on a local target. (#15921, #16040, #16032)

🏗️ Core Agent & Model Runtime
- Rework input completion prompt to v1.2 — predict intent, including long-range. (#15961)
- Inject model knowledge cutoff into context. (#16048)
- Support GLM-5.2 reasoning effort. (#15972)
- Derive OpenAI model routing from IDs. (#16046)
- Rename model type stt to standard asr (runtime migration, non-breaking). (#16002)
- Persist model config cache for faster startup. (#16047)
- Hetero-agent: assistant-anchored message-chain write spine. (#15930)
- Stop chat mode from auto-injecting local-system tools. (#15981)
- Centralize inbox agent meta fallback across read paths. (#15948)
- Expose missing usage diagnostics. (#15973)

🎙️ Audio & ASR
- Add ASR (transcribe) with OpenAI + Gemini-native backends, exposed via tRPC. (#15992)
- Pass audio (mp3) attachments to audio-capable models like Gemini. (#15988)
- Allow audio upload in agent mode and render audio with a waveform player. (#16022)
-

This release was automatically published from PR #15947.
Changes
See PR description: https://github.com/lobehub/lobehub/pull/15947
Commit Message
🚀 LobeHub Release (20260617)
Release Date: June 17, 2026
Since v2.2.5: 42 commits · 42 merged PRs · 8 contributors
This weekly release deepens server-side agent orchestration, brings desktop file and worktree capabilities to the web through device RPC, and smooths out the everyday rough edges — cold-start boot, connector credential safety, and chat refresh feedback.

✨ Highlights
- Server-side Group Orchestration — Agents can now call other agent members server-side, enabling multi-agent collaboration without a desktop in the loop. (#15870)
- Desktop File Ops on the Web — Project file operations and git worktree listing now replicate from desktop to web via device RPC, so cloud sessions can read and act on local working directories. (#15885, #15889)
- Fleet Running-Tasks Dashboard — A lab-gated dashboard for in-flight Fleet tasks, with running topics re-synced each time the Observation tab opens. (#15817, #15922)
- callAgent as a Deferred Tool — The agent loop now runs callAgent as a deferred tool, giving cleaner sub-agent invocation and tool-chain handling. (#15765)
- Connector Credential Safety — Editing a connector no longer risks silently wiping saved credentials; they are restored in edit mode and preserved on save. (#15909)
- Smoother Cold Start — Boot now shows one continuous loading screen instead of a brand-logo flash on cold start. (#15926)

🏗️ Core Agent & Architecture
- Improved connector, document, and Fleet agent workflows. (#15936)
- Scoped the agent conversation subtree to an explicit agentId for clearer multi-agent boundaries. (#15866)
- Added a role-aware dual-form message-chain reader to the conversation flow. (#15908)
- Anchored the server-side main chain to a run's real last tool in heterogeneous agents. (#15883)
- Drove resume completion off the authoritative Durable Object status in the gateway client. (#15919)
- Corrected target agentId and refreshed the sidebar in gateway mode for the agent builder. (#15888)
- Forwarded model extend params on the server-side agent runtime. (#15891)
- Preserved preference-memory receipt routing in agent signals. (#15892)
- Filtered the .tool-results archive out of document lists by default. (#15935)
- Optimized the agent document list query. (#15904)

📱 Devices & Platforms
- Locked a run to the explicitly selected device, never offering device-switching mid-run. (#15914)
- Exposed deviceRouter on the mobile router. (#15925)
- Opened a new Home tab from the desktop tab bar "+" button. (#15825)
- Added support for approved external local file previews on desktop. (#15895)
- Removed web onboarding aliases from the desktop build. (#15902)
- Consolidated auth SPA loading. (#15903)

🖥️ User Experience
- Fixed assorted workspace problems and clarified workspace copy/move actions. (#15928, #15897)
- Showed a cached message-refresh hint with breathing room around it. (#15901, #15906)
- Added an unread-reply indicator on collapsed project groups. (#15915)
- Anchored the sidebar spacer immediately after the accordion block. (#15871)
- Capped nested thread-list height with scroll overflow. (#15861)
- Removed the ParamsPanelToggle control icon from the chat header. (#15860)
- Defaulted the React Scan scanner UI to off. (#15934)
- Refined top-up best-value and referral reward-rules copy. (#15924, #15923)
- Only enforced the chat upload file-type whitelist in chat mode. (#15884)

🔒 Reliability
- Deduped unread-count polling to reduce redundant requests. (#15881)
- Aligned dayjs locale imports. (#15896)

👥 Contributors
Huge thanks to the 8 contributors who shipped 42 merged PRs across 42 commits this cycle.
@arvinxx · @Innei · @tjx666 · @LiJian · @AmAzing129 · @sudongyuer · @rivertwilight · @Rdmclin2

Full Changelog: https://github.com/lobehub/lobehub/compare/v2.2.5...release/weekly-20260617

This release was automatically published from PR #15877.
Changes
See PR description: https://github.com/lobehub/lobehub/pull/15877
Commit Message
🚀 LobeHub Release (20260615)
Release Date: June 15, 2026
Since v2.2.4: 48 merged PRs · 5 contributors
This cycle lands the Composio integration as the new connector backbone, a unified tiered client cache, and a deep round of agent-runtime reliability hardening for cold-replica and sub-agent flows.

✨ Highlights
- Composio integration — New Composio integration layer replaces Klavis as the connector backbone for third-party skills. (#15461)
- Tiered client cache — Unified localStorage + IndexedDB cache provider with per-scope isolation, plus a registry-wide convergence of SWR keys for predictable invalidation. (#15844)
- Gateway mode in chat config — Gateway mode now lives in chat config, making it per-conversation rather than a global toggle. (#15714)
- Bulk move topics — Move multiple topics to another assistant in one action. (#15809)
- Skills row actions — View / rename / delete row actions in the working sidebar, plus edit / uninstall for connectors in Skill detail. (#15864, #15829)
- Token usage cache rate — Conversations now surface the prompt-cache hit rate alongside token usage. (#15812)

🏗️ Core Agent & Architecture
- Run lifecycle — Extracted client run-completion into a shared buildRunLifecycle, with a characterization net over agent-runtime run-lifecycle. (#15854, #15843)
- Sub-agent resilience — Hardened async sub-agent suspend/resume against missed wakeups. (#15855)
- Cold-replica correctness — Fixed main-turn idempotency and now mark topics failed on terminal errors; persist sub-agent turn id so cold replicas don't fragment a turn; dedupe sub-agent thread creation after finalize. (#15838, #15808, #15849)
- Stream routing — Drop sub-agent-tagged events from the main gateway stream handler, and preserve subAgentId / documentId in the message bucket key context. (#15814, #15865)
- Heterogeneous agents — Forward bot / IM image attachments to heterogeneous agents. (#15868)
- Agent state — Stop background config fetch from hijacking the active agent, and warn when agent mode is on but the model lacks tool calling. (#15862, #15828)
- Tracing — Enable S3 tracing by default in production. (#15841)

🔌 Integrations & Skills
- Skill panel — Dedupe skill-panel rows and allow deleting pending integrations; stop connected integrations from duplicating in the chat-input skill panel. (#15872, #15869)
- Connectors — Edit / uninstall buttons for connectors in Skill detail. (#15829)

🖥️ Chat & User Experience
- Topics — Server-side status filter via a new queryTopics query, and per-agent topic search scoped by agentId. (#15822, #15798)
- Message rendering — Render mixed assistant blocks in natural order, fold short mixed tool blocks together, and render mention names from the serialized attribute instead of falling back to "unknown". (#15810, #15857, #15831)
- Tool workflow — Tool-workflow collapse no longer shows "in progress" once content renders below it. (#15815)
- Token usage — Derive operation token usage from messages rather than a parallel accumulation. (#15819)
- Reconnect — Normalize reconnect startTime to epoch ms. (#15811)
- Home & editor — Hide the agent-mode notice while config is loading, and isolate the page-editor copilot context from global agent/document state. (#15846, #15826)
- Polish — base-ui modal fixes the provider delete-confirm z-index, the updater renders release notes as Markdown, revert-confirm and toast copy tightened. (#15845, #15867, #15813)
- Desktop — Tray double-click opens the main window. (#15816)

🔒 Reliability
- Auth gating — Gate the listDevices request behind login state so it no longer fires before authentication. (#15876)

🔧 Tooling & Internal
- SWR convergence — Converged store-, UI-, and straggler SWR keys into the swrKeys registry, fixing a stale prefetch key along the way. (#15863, #15858, #15853, #15850, #15848)
- Tests — Characteriz

This release was automatically published from PR #15806.
Changes
See PR description: https://github.com/lobehub/lobehub/pull/15806
Commit Message
🚀 LobeHub Release (20260614)
Release Date: June 14, 2026
Since v2.2.3: 99 commits · 99 merged PRs · 11 contributors
This cycle deepens cross-device collaboration — browser pairing, a shared desktop/CLI device gateway, and edit locks that keep multiple agents and people aligned on the same Context.

✨ Highlights
- Browser device pairing — Pair a browser as a device and route agent tools to it, with rename/delete actions on the branch switcher. (#15678, #15774)
- Shared device gateway — Desktop and CLI now share one remote-device gateway RPC, so device-bound runs behave the same everywhere. (#15780)
- Operation status tray — A live op-status tray sits above the chat input, tracking operation usage and staying compact on narrow screens. (#14737, #15736, #15735)
- Inline file previews — HTML files render inline and remote read-only local files preview directly in the portal. (#15671, #15673)
- New providers — Added AntGroup (蚂蚁百灵), Longcat with live model-list fetch, and new SenseNova models. (#13713, #15134, #15306)
- Desktop tab management — Drag-to-reorder desktop tabs, plus restored cloud desktop builds. (#15787, #15666)

🏗️ Core Agent & Runtime
- Heterogeneous chaining — Stabilized main-message chaining and unified the client hetero executor on a shared mainAgentReducer. (#15783, #15762)
- Sub-agent resilience — Block recursive server sub-agents, keep async sub-agent streams alive, and rehydrate sub-agent runs from DB on cold replicas. (#15731, #15646, #15788)
- Reasoning persistence — Always persist assistant reasoning to the DB so it survives reloads. (#15687, #15690)
- Device routing — Resolve device routing and device-tool injection through a single execution plan. (#15669, #15683)
- Image attachments — Persist and deliver image attachments for device/sandbox hetero runs. (#15685)
- Virtual sub-agents — Split the virtual sub-agent entry and clarified its naming. (#15733, #15737)

🖥️ Chat & User Experience
- Topic management — Topic sidebar status indicators, selector topic actions, and a batchMoveTopics mutation for bulk moves. (#15739, #15744, #15793)
- Local file portals — Scope local file tabs by working directory and auto-close empty local previews. (#15732, #15760)
- Editing — Coalesce document autosave history into 10-minute windows and fold connector OAuth into the custom MCP form. (#15716, #15661)
- Skills — Delete/remove actions on settings skill items. (#15708)
- Polish — Preserve message order after tool results and stop ContentLoading from leaking raw operation i18n keys. (#15657, #15752)

🤖 Models & Providers
- Model bank metadata — knowledgeCutoff batch 2 with a metadata skill and an always-visible tab bar, plus backfilled family/generation data. (#15663, #15642, #15640)
- Provider quality — Improved DeepSeek structured output, Kimi code thinking mode, and a model guard kept in provider grouping. (#15680, #15725, #15681)
- Discoverability — Surface model-list fetch failures instead of failing silently. (#15753)

🔒 Reliability & Security
- Error classification — Classify "Agent state not found" as StateStoreReadError, classify untyped Error throws via message patterns, and surface missing tool calls as errors. (#15778, #15767, #15691)
- Codex — Parse retry time in the stated timezone and detect the bundled Codex CLI from Codex.app on macOS. (#15758, #15759)
- Mobile — Stop the pushToken.unregister 401 storm while preserving authenticated legacy cleanup, and gate inbox unread count by login state. (#15719, #15723, #15724)
- Performance — Derive topic activity from messages and drop sitemap generation to cut static export time. (#15726, #15702)
- Security: Bumped @opentelemetry/auto-instrumentations-node, @opentelemetry/sdk-node, and vitest. (#14686, #14687, #15698)

🔧 Tooling & Docs
- Agent testing — Merged local-testing and cli-backend-testing into a single ag

This release was automatically published from PR #15647.
Changes
See PR description: https://github.com/lobehub/lobehub/pull/15647
Commit Message
🚀 LobeHub Release (20260610)
Release Date: June 10, 2026
Since v2.2.2: 131 merged PRs · 13 contributors
This weekly release strengthens agent collaboration across cloud, desktop, CLI, and workspace flows, with steadier runtime behavior and a broader foundation for workspace-scoped data.

✨ Highlights
- Agent execution across devices — Unifies per-device working directories, project skill discovery, and sub-agent suspend/resume behavior across server, QStash, and device RPC flows. (#15543, #15566, #15481, #15620, #15591)
- Connector and sandbox platform — Expands connector permissions, custom OAuth MCP connector onboarding, sandbox provider support, and user-uploaded file sync into cloud sandbox runs. (#15463, #15546, #15184, #15550)
- Desktop and CLI reliability — Fixes desktop cold-start, auto-update, Windows build, CLI skill discovery, and lh connect agent dispatch paths. (#15547, #15525, #15527, #15562, #15632, #15634)
- Pages and sharing — Refreshes topic sharing, improves Page Editor layout behavior, and routes Page Agent tool execution through the server-side editor path. (#15581, #15556, #15588, #15023, #15610)
- Model availability and provider updates — Adds user-scoped LobeHub model availability, Claude Fable 5, Qwen thinking preservation, and MiniMax M3 updates. (#15590, #15639, #13494, #15376)

🏗️ Core Product & Architecture
Agent Runtime & Heterogeneous Agents
- Improves sub-agent lifecycle handling, including async suspend/resume, queue-mode QStash resume delivery, and blocking nested sub-agent calls. (#15481, #15620, #15575)
- Stabilizes heterogeneous agent ingestion and streaming with raw stream dumps, per-turn usage, image forwarding on regenerate, and duplicate-text fixes. (#15602, #15577, #15592, #15585)
- Adds execution-device and working-directory controls across device RPC, legacy defaults, and remote-spawned Claude Code sessions. (#15543, #15566, #15591, #15572)
- Improves runtime diagnostics and compatibility, including Gemini multimodal output capture, abort stream semantics, and trace quality analysis. (#15535, #13677, #15508)

📱 Platforms, Integrations & UX
Connectors, Sandbox & Tools
- Ships API-level connector tool permissions, custom OAuth MCP connector onboarding, and connector-first runtime execution. (#15463, #15546)
- Adds sandbox provider support, cloud sandbox file sync, and safer external URL file input handling with SSRF validation. (#15184, #15550, #12657)
- Improves tool visibility and execution with pinned app-fixed tools, ANSI output rendering, gateway-tunneled MCP calls, and automatic headless tool runs. (#15509, #15516, #15469, #15492)
Desktop, CLI & Web UX
- Restores desktop startup and reload behavior, preserves IPC error causes, and keeps the tab bar new-tab action visible across routes. (#15547, #15597, #15638)
- Fixes desktop update and build stability for browser quit guards, macOS update signing, and Windows Visual Studio detection. (#15525, #15527, #15562)
- Shows the plan-limit upgrade UI on desktop builds. (#15628)
- Adds the Agent Run delivery checker and fixes CLI device dispatch plus skill list/search output. (#15489, #15634, #15632)
- Refreshes onboarding, auth source preservation, topic UI states, referral/Fable campaign copy, and chat-input control bar behavior. (#15629, #15544, #15573, #15614, #15616, #15617, #15622, #15643)

🔒 Security, Reliability & Rollout Notes
- External URL file input now includes SSRF validation for safer Google file handling. (#12657)
- Database workspace-scope migrations are part of this release; self-hosted operators should run the normal migration path before serving the updated app. (#15446, #15465, #15468, #15472)
- The release branch was re-cut from canary and includes the latest main release-version commit so v2.2.2 is the verified compare base.

👥 Contributors
@ONLY-yours, @sxjeru, @hard

This release was automatically published from PR #15447.
Changes
See PR description: https://github.com/lobehub/lobehub/pull/15447
Commit Message
🚀 LobeHub Release (20260604)
Release Date: June 4, 2026
Since v2.2.1: 88 merged PRs · 11 contributors
This week brings Execution Devices out of the lab — run agents and Claude Code on any configured local or remote machine — alongside Claude Opus 4.8, token-usage analytics, and Page sharing.

✨ Highlights
- Execution Devices — Pick where an agent runs. Desktop and CLI devices auto-register with a stable machine ID, route through the gateway by channel, and surface a device switcher in the chat input. Run remote Claude Code on a configured device, with a recent-directory picker you can drag to reorder. (#15300, #15315, #15322, #15343, #15351, #15371)
- Claude Opus 4.8 — Day-one support for Anthropic's latest model. (#15314)
- Token-usage analytics — A new token-usage mode on the activity heatmap, backed by a denormalized topic usage/cost rollup so totals stay accurate without recomputing from messages. (#15365, #15417, #15425)
- Page sharing — Share a Page through a dedicated document share flow, plus new Workspace and Agent share tables. (#15309, #15439)
- Self-iteration agents — Agent Signal's execAgent migration lands a server-runtime bridge, async memory writer, and a registered self-iteration tool package, with a CLI trigger command for testing. (#15360, #15364, #15392)
- Knowledge search — BM25 search now extends to file-backed documents, and the portal ships an editable CodeMirror viewer for local files with document highlighting. (#15247, #15298)

🏗️ Core Agent & Architecture
Agent Signal & Runtime
- execAgent migration — Server-runtime bridge, completion projection, async memory writer, and removal of the legacy executeSelfIteration path. (#15392)
- Registered the self-iteration builtin tool package and restored the three mode-specific self-iteration agent slugs. (#15202, #15364)
- Added a CLI trigger command with a golden-snapshot fixture for Agent Signal. (#15360)
- Skill priority — Agent Builder now emits a skill-priority instruction with matching server runtime. (#15409)
- Retry empty LLM completions instead of silently finishing the turn. (#15355)
- Classify topic/agent/session foreign-key violations as ConversationParentMissing for clearer recovery. (#15408)
- Persist canonical nested usage/performance on assistant messages, and re-link orphan tool messages at the raw bucket write boundary. (#15359, #15438)
- Guard createAgent against LLM double-encoded array fields. (#15381)

🖥️ Execution Devices & Gateway
- Auto-register desktop and CLI devices with a stable machine ID, and add the @lobechat/device-identity package. (#15300, #15321)
- New Devices settings page behind the Execution Device Switcher lab, with a device switcher shown for all agents in the chat input. (#15315, #15371)
- connectionId + channel routing across the gateway client and device list; preset the local device on the first LLM request for the 本机 target. (#15322, #15435)
- Run remote Claude Code on a configured device, with drag-to-reorder recent-directory management and client renders for device tool results. (#15343, #15351, #15437)
- Preserve content and state across gateway tool calls, and prevent duplicate streaming from stale reconnects. (#15114, #15354)

🖥️ CLI & Desktop
- Preserve content/state for connect local file and shell tools; render the runCommand tool result card. (#15441, #15442)
- New lh topic view command; CLI now auto-registers its device on login, matching desktop. (#15340, #15377)
- Resolve CLI tools from the shell PATH, and clarify local command session handling. (#15368, #15389)
- Relocate visual-ref helpers to @lobechat/const to fix a renderer crash; upload .blockmap files to S3 for differential updates. (#15326, #15369)
- Fix a market OAuth expiry that triggered the wrong re-login modal, and kill dev child processes on parent shutdown. (#15246, #15290)

🗂️ Pages, Librar

This release was automatically published from PR #15302.
Changes
See PR description: https://github.com/lobehub/lobehub/pull/15302
Commit Message
🚀 LobeHub Release (20260528)
Release Date: May 28, 2026
Since v2.2.0: 220 merged PRs · 15 contributors
This cycle brings heterogeneous "platform agents" you can dispatch to local or remote devices, a rebuilt onboarding flow, document-centric chat, and a unified model-runtime error model — with new DeepSeek V4 and Gemini 3.5 Flash support along the way.

✨ Highlights
- More Hetero Agents (OpenClaw / Hermes) — Create heterogeneous agents and dispatch them to local or remote devices through the device gateway, with an execution-target switcher in the composer and persistent CLI sessions. (#15065, #15179, #15022)
- iMessage on Desktop — New iMessage setup and bridge on desktop, plus bot attachments across every platform. (#15228, #15227, #15029)
- Skills in the Composer — Drag skill chips into chat, trigger installed skills from the slash menu mid-line, and surface project-level skills in the homogeneous agent runtime. (#15095, #15061, #15110)
- New Models — DeepSeek V4 Flash/Pro and Gemini 3.5 Flash across providers, with thinking params for structured output and chat cost estimates. (#15031, #15001, #15051, #14876)
- Agent Runtime Observability — OpenTelemetry GenAI semantic conventions plus per-call generation tracing. (#15123, #15124)

🤖 Agents & Heterogeneous Runtime
- Platform agent creation — OpenClaw/Hermes creation UI, device guard, and remote dispatch backend. (#15065)
- Execution-target switcher — Pick local vs remote execution directly in the composer; device-selection UX with actionable guidance. (#15179, #15111)
- CLI hetero dispatch — OpenClaw/Hermes dispatch with persistent sessions and a notify protocol. (#15022)
- Gateway snapshot as source of truth — Consume the gateway uiMessages snapshot at step boundaries to keep chat state consistent. (#15153, #15152)
- Client sub-agent as a normal tool call — Simplifies the sub-agent execution path. (#15281)
- Hermes agent chain — Implements the Hermes agent chain logic. (#15189)
- Device registry — TRPC endpoints to register, list, update, and remove devices. (#15299)
- Desktop device routing — Route gateway agent runs through lh hetero exec; restore userId in gateway dispatch and gate local-system by execution target. (#15132, #15232)
- Agent signals — Anchor agent-signal receipts to messages and isolate memory-agent messages into a child thread. (#14969, #14921)

🚀 Onboarding
- Simplified first screen — Defer topic creation to first send. (#15090)
- Market Agent Picker — Added as a classic onboarding step, with template prefetch. (#14980, #15041)
- Welcome guidance — Show agent welcome guidance on first run. (#15098)
- Mobile — Adapt agent onboarding UI and restore Classic-step padding on mobile. (#15019, #15032)
- Discovery — Streamline discovery to a single profession question. (#14987)
- Analytics — Track onboarding step events and create-agent modal source. (#15133, #15028)

📄 Documents, Pages & Knowledge
- Thread chat in preview — Embed thread chat in the document preview portal. (#15216)
- Non-markdown rendering — Render non-markdown docs as a read-only highlight. (#15272)
- Multi-select — Multi-select delete in the document tree. (#15125)
- Page-agent streaming — Preview initPage streaming arguments. (#15039)
- Per-agent topics — Per-agent topic management page. (#15207)
- Server-side category — Derive document category server-side and drop frontend predicates. (#15076)

🧩 Skills & Tools
- Drag skill chips — Drag skills into chat input and register agent-document skills. (#15095)
- Slash menu — Installed skills appear in the slash menu with a mid-line trigger. (#15061)
- Project skills — Recognize project-level skills in the homogeneous agent runtime and surface them regardless of active device. (#15110, #15177)
- VFS archiving — Archive oversized tool results to VFS instead of truncating. (#15074)
- @localFile ment

This release was automatically published from PR #14739.
Changes
See PR description: https://github.com/lobehub/lobehub/pull/14739
Commit Message
🚀 LobeHub Release (20260513)
Hotfix Scope: Ship the canary backlog (111 PRs) onto main as a fast-tracked patch — operator-focused, no weekly-style write-up.
Brings the accumulated canary work into main: agent/task improvements, hetero-agent fixes, desktop & onboarding polish, and several reliability caps.
✨ What's Included
- Agent & tasks — Self-review proposal-to-action automation, sub-agent dispatch consolidated to lobe-agent, AskUserQuestion wiring for Claude Code, scheduler/hotkey/TodoList polish. (#14583, #14657, #14715, #14639, #14732, #14707, #14713)
- Home & onboarding — Daily brief with linkable welcome + paired input hint, inline skill auth in recommended task templates, cleanup of captcha-on-signin and marketplace early-exit. (#14589, #14676, #14573, #14598)
- Bots & integrations — Slack MPIM support, Discord DM fix, slash-command + connect-error fixes, gateway client-tool plugin state. (#14733, #14591, #14596)
- Desktop & CLI — Windows .cmd shim detection for claude / codex CLIs, auth focus & pending-login reset fixes. (#14720, #14694, #14695)
- Reliability — Cap web-crawler body size and image binary at safe limits, attach error listeners to Neon/Node pools, reject inactive OIDC access. (#14660, #14711, #14606, #14674)
- Database — agent_operations table + persist agent operations from the runtime; switch user memory search to paradedb.match(...). (#14416, #14736, #14590)
⚙️ Upgrade
- Self-hosted: pull the latest image and restart. Drizzle migrations (including the new agent_operations table) run automatically on boot.
- Cloud: applied automatically on deploy.

Full Changelog: https://github.com/lobehub/lobehub/compare/v2.1.57...hotfix/v2.1.57-ce08b9b1

This release was automatically published from PR #14354.
Changes
See PR description: https://github.com/lobehub/lobehub/pull/14354
Commit Message
🚀 LobeHub v2.1.56 (20260501)
Release Date: May 1, 2026
Migration Scope: briefs table — 2 new columns (trigger, metadata), 1 new index (briefs_trigger_idx)
🗄️ Migration Overview
File: packages/database/migrations/0100_add_metadata_and_trigger_to_briefs.sql
Added columns on briefs:
- trigger varchar(255) — module that triggered the brief (e.g. task, agent, signal)
- metadata jsonb — freeform business state attached by the producer
Added index:
- briefs_trigger_idx on briefs(trigger)
The migration uses ADD COLUMN IF NOT EXISTS and CREATE INDEX IF NOT EXISTS, so re-running it is safe.

⚙️ Operator Notes
- Migration runs automatically on application startup.
- Both new columns are nullable — existing rows are unaffected and no backfill is required.
- The migration is additive only; no existing column or index is modified or dropped.
- Recommended: take a routine snapshot before deployment, as with any schema change.

🔒 Reliability & Risk
- Backwards-compatible: Older application code that does not read or write trigger / metadata continues to work, since both columns are nullable.
- Index cost: briefs_trigger_idx is built with IF NOT EXISTS and is small (single varchar(255) column); build time on existing data is expected to be negligible.
- Rollback: If a rollback is required, drop briefs_trigger_idx and the two new columns; no data migration is involved.

👥 Owner
Migration owner: @nekomeowww
The migration owner is responsible for rollout follow-up and incident handling for this schema change.

This release was automatically published from PR #14284.
Changes
See PR description: https://github.com/lobehub/lobehub/pull/14284
Commit Message
Hotfix Scope: Topic preservation across cold chat-entry routes
Keeps newly created Topics visible when a first message is sent before the destination chat route has fully hydrated.
🐛 What's Fixed
- Page Agent empty-session regression — Sending the first message in an empty Page Agent panel no longer clears the newly created Topic and returns the panel to an empty state. (Resolves LOBE-8351)
- Home cold-route send regression — Sending from the Home default Chat Input now routes to the newly created Inbox Topic even when /agent/:aid has never been opened and the route chunk has no warm cache.
- Page-scoped Copilot consistency — Page Copilot and File Copilot share the same provider-level topic reset behavior, so stale Topics are cleared only when entering or switching the scoped Agent.
- Regression coverage — Added focused unit coverage for Home default sends, route parity coverage remains intact, and added an E2E scenario for the no-cache Home send path.
✅ Verification
- bunx vitest run --silent='passed-only' 'src/routes/(main)/home/features/InputArea/useSend.test.ts' 'src/spa/router/desktopRouter.sync.test.tsx' 'src/routes/(main)/agent/features/Conversation/ChatHydration/index.test.tsx' 'src/routes/(main)/agent/_layout/AgentIdSync.test.tsx'
- BASE_URL=http://localhost:3007 DATABASE_URL=postgresql://postgres:postgres@localhost:5433/postgres bun run test -- --tags '@HOME-CHAT-COLD-001' from e2e/
⚙️ Upgrade
- Self-hosted: pull the new image and restart. No schema or environment changes.
- Cloud: ships through the normal hotfix deployment after merge.
👥 Owner
@Innei
Fixes LOBE-8351

This release was automatically published from PR #14231.
Changes
See PR description: https://github.com/lobehub/lobehub/pull/14231
Commit Message
Hotfix Scope: Agent topic / thread navigation regression — stale chat state on agent switch
Clears residual topic state when navigating between agents, restores the active subtopic's title in the header, and keeps the sidebar's thread list expanded while a thread is open.
🐛 What's Fixed
- Stale topic on agent switch — ChatHydration syncs activeTopicId / activeThreadId from the URL via useLayoutEffect and writes null (not undefined) so /agent/agt_A/tpc_X → /agent/agt_B no longer carries over the previous topic; Start new topic responds again.
- Conversation context isolation — ConversationProvider keys its inner store on contextKey, so consumers don't read stale values for one render after agent / topic / thread identity changes.
- Sidebar thread list visibility — <ThreadList /> visibility is now driven by urlTopicId and accepts topicId as a prop, so the parent topic's thread list stays expanded while viewing a subtopic.
- Header thread title — Header Tags reads the active thread's title from s.threadMaps[s.activeTopicId] when activeThreadId is set, falling back to chat:thread.title for unnamed threads.

This release was automatically published from PR #14217.
Changes
See PR description: https://github.com/lobehub/lobehub/pull/14217
Commit Message
🚀 LobeHub v2.1.53 (20260427)
Release Date: April 27, 2026
Since v2.1.52: 194 merged PRs · 17 contributors
Introduce Heterogeneous Agent — Claude Code and Codex run as first-class desktop runtimes, paired with a new Agent Signal package, sharper desktop UX, and a wave of flagship model additions.

✨ Highlights
- Introduce Heterogeneous Agent — Claude Code and Codex run as first-class desktop agents: subagent rendering, partial-message streaming, multi-turn resume, terminal error surfacing, rich tool inspectors, and runtime polish. (#14162, #13754, #14067, #14001, #13970, #13942)
- Screen capture & Quick Chat tray — New desktop screen capture overlay (macOS permission-gated) with Quick Chat tray and upload pipeline improvements; chat input auto-focuses on overlay mount. (#13818, #14097, #14105)
- Desktop topic & tab UX — Dedicated topic popup window with cross-window sync, Cmd+W/Cmd+T tab shortcuts, TabBar polish, recent working directories expanded to 20, and human approval notifications. (#13957, #13983, #13972, #14036, #14092)
- Git workflow built-in — One-click pull/push from the branch chip, ahead/behind badge, and submodule/worktree repo detection. (#14041, #13980, #13978)
- Agent Signal package — New @lobechat/agent-signal runtime for dynamic memory feedback signals, with OTel metrics and self-iteration in Lab. (#14157, #14170, #14159, #14169, #14187)
- New models — Claude Opus 4.7 with xhigh effort tier, GPT-5.5, DeepSeek V4 Flash/Pro with reasoning slider, Kimi K2.6, MiMo-V2.5/Pro, gpt-image-2, Qwen3.6 Flash/Plus, and Pixverse-c1. (#13903, #14147, #14114, #14004, #14089, #14039, #13923)
- New providers — OpenCode Zen, OpenCode Go, and Azure OpenAI Router runtime. (#13943, #14064, #13823)
- Mobile settings overhaul — Full settings menu and responsive profile layout for mobile. (#14019)

🏗️ Heterogeneous Agent
- Claude Code runtime, working-directory awareness, and sidebar polish. (#13970)
- CC subagent rendering with persistent streamed text; parallel-tool orphan fix. (#14001, #13968, #14024)
- Per-step usage persisted to each step assistant message. (#13964)
- Per-phase workflow expand defaults; full-expand toggle with three-level expansion. (#14171, #13906)
- Hetero-mode actions bar; tool inspector polish. (#13963, #14034, #14030)
- Codex desktop integration with rich tool rendering and devtools preview. (#14067, #14100)
- Codex terminal error surfacing and CLI output tracing. (#14166)
- Tighten isCanUseVision default and add aggregator fallback. (#14172)
- Persist ccSessionId in topic metadata for CC multi-turn resume. (#13902)
- CC account card, topic filter, and integration polish. (#13955, #13942, #13950)
- Token-level deltas streamed via --include-partial-messages. (#13929)

🧠 Agent Signal & Self-Iteration
- New @lobechat/agent-signal package with dynamic feedback signals. (#14157)
- AgentSignalRuntime wired through agent-tracing and observability-otel metrics. (#14170, #14159)
- Self-iteration feature flag added to Lab; front-side flag check. (#14169, #14186)
- Signal policy for receiving memory feedback dynamically. (#14187)

💬 Conversation
- Queue follow-up sends during running CC turns. (#14179)
- Persist per-topic chat scroll position; pin user message + fold long messages. (#14191, #14056)
- Inline resend when editing last user message. (#14080)
- Disable first-block markdown streaming to prevent flicker. (#14193, #13904)
- Prevent Markdown stream replay when vlist remounts streaming items. (#14086)
- Stop repinning after manual scroll; unify scroll-to-user + spacer hooks. (#14099, #14132)

📱 Platforms & Integrations
Desktop / Electron
- Screen capture overlay, Quick Chat tray, and upload pipeline improvements. (#13818)
- macOS permission gate for screen capture; auto-focus chat panel input. (#14097, #14105)
- Dedicated topic popup window with cross-window syn

This release was automatically published from PR #13994.
Changes
See PR description: https://github.com/lobehub/lobehub/pull/13994
Commit Message
This release ships a database schema migration that makes two pieces of lifecycle state explicit: how a topic finishes, and how a task is automated. Both already existed implicitly, which made some downstream behaviors ambiguous. This change turns them into first-class columns so the UI, scheduler, and reporting layers can rely on them directly.
Migration overview
Previously, a topic's completion was inferred from surrounding context — there was no dedicated status or timestamp marking when a topic moved out of the active pool. Tasks had a similar ambiguity around automation: heartbeat supervision and cron-style scheduling lived in overlapping fields, so a configured-but-paused schedule looked identical to one that was never set.
This migration makes both states explicit.
topics
- status (active | completed | archived)
- completed_at timestamp
- Indexes on status and (user_id, completed_at) so completion views and archival queries stay cheap
tasks
- automation_mode (heartbeat | schedule | null) — a single discriminator that makes the two automation kinds mutually exclusive by design and removes the "configured but disabled" ambiguity
- Index on automation_mode
In practice, the topic list can now distinguish completed and archived records without reinterpreting other fields, and the task scheduler gets a clean switch to route records between heartbeat supervision and cron execution.
Notes for self-hosted users
- The migration runs automatically during app startup; no manual SQL action is required in standard deployments.
- All new columns are nullable, so existing rows stay untouched — nothing is backfilled at migration time.
- As with any schema release, we still recommend a database backup and rollout during a low-traffic window.
The migration owner: @arvinxx — responsible for this database schema change, reach out for any migration-related issues.

This release was automatically published from PR #13895.
Changes
See PR description: https://github.com/lobehub/lobehub/pull/13895
Commit Message
Release Date: April 16, 2026
Since v2.1.49: 107 commits · 101 merged PRs · 13 contributors
This weekly release focuses on improving runtime stability and gateway execution consistency, while making Home/Recents workflows faster to navigate and easier to manage in daily use.

✨ Highlights
- Server-side Human Approval Flow — Agent runtime now supports more reliable approve/reject/reject-continue handling in gateway mode, reducing stalled execution paths in long-running tasks. (#13829, #13863, #13873)
- Message Gateway End-to-End Hardening — Gateway message flow, queue handling, tool callback routing, and stop interruption behavior were strengthened for better execution continuity. (#13761, #13816, #13820, #13815)
- Client Tool Execution in Gateway Mode — Client-executor tools now run more predictably across gateway and desktop callers, with improved executor dispatch behavior. (#13792, #13790)
- Home / Recents / Sidebar Upgrade — Sidebar layout, custom sort, recents operations, and profile actions were improved to reduce navigation friction in active sessions. (#13719, #13812, #13723, #13739, #13878, #13734)
- Agent Workspace and Documents Expansion — Working panel and agent document workflows were expanded and polished for better day-to-day agent operations. (#13766, #13857)
- Provider and Model Compatibility Improvements — Added GLM-5.1 support and refined model/provider edge-case handling, including schema and error-path fixes. (#13757, #13806, #13736, #13740)

🏗️ Core Agent & Architecture
Agent runtime and intervention lifecycle
- Added server-side human approval and improved runtime coordination across approve/reject decision paths. (#13829, #13863)
- Improved interrupted-task handling and operation lifecycle consistency to reduce half-finished runtime states. (#13714)
- Refined error classification and payload propagation so downstream surfaces receive clearer actionable errors. (#13736, #13740)
Execution model and dispatch behavior
- Introduced executor-aware runtime behavior to better separate client/server tool execution semantics. (#13758)
- Improved tool/plugin resolution and manifest handling to avoid runtime failures on malformed inputs. (#13856, #13840, #13807)

📱 Gateway & Platform Integrations
- Added message gateway support and strengthened queue/error behavior for more stable cross-channel execution. (#13761, #13816, #13820)
- Improved gateway callback pipeline with protocol and API additions for tool_execute / tool_result. (#13762, #13764, #13765)
- Improved bot/channel reliability and DM/slash handling in Discord-related paths. (#13805, #13724)

🖥️ CLI & User Experience
- Improved CLI reliability across message/topic operations and build/minify-related paths. (#13731, #13888)
- Added image-to-video options and improved command behavior for generation workflows. (#13788)
- Improved desktop runtime behavior for remote fetch and Linux notification urgency handling. (#13789, #13782)

🔧 Tooling
- Extracted gateway stream client into @lobechat/agent-gateway-client to centralize protocol usage and reduce duplication. (#13866)
- Improved built-in tool coverage and runtime support, including GTD server runtime and missing lobe-kb tools. (#13854, #13876)
- Updated skill and frontmatter consistency in workflow tooling. (#13730)

🔒 Security & Reliability
- Security: Strengthened API key WS auth behavior and safer serverUrl forwarding in gateway-related auth paths. (#13824)
- Reliability: Reduced runtime stalls by improving gateway stop/interrupt and approval-state routing behavior. (#13815, #13863, #13873)
- Reliability: Added defensive guards for malformed tool manifests and non-string content edge cases. (#13856, #13753)

👥 Contributors
101 merged PRs from 13 contributors across 107 commits.
Community Contributors
- @arvinxx - Runtime, gateway, and execution reliability

This release was automatically published from PR #13789.
Changes
See PR description: https://github.com/lobehub/lobehub/pull/13789
Commit Message
💻 Change Type
-
✨ feat
-
🐛 fix
-
♻️ refactor
-
💄 style
-
👷 build
-
⚡️ perf
-
✅ test
-
📝 docs
-
🔨 chore
🔗 Related Issue
- None
🔀 Description of Change
- Extract the document history database changes from the feature branch onto a branch based on main.
- Add the document history migration, schema, relations, model, and database tests only.
- Exclude UI, router, and service-layer changes so the PR stays focused on the database layer.
🧪 How to Test
- Run: cd packages/database && bunx vitest run --silent=passed-only src/models/tests/document.test.ts src/models/tests/documentHistory.test.ts
-
Tested locally
-
Added or updated tests
-
No tests needed
📸 Screenshots / Videos
Before After
N/A N/A
📝 Additional Information
- This PR intentionally targets main because the database migration needs to land on the release branch first.

This release was automatically published from PR #13716.
Changes
See PR description: https://github.com/lobehub/lobehub/pull/13716
Commit Message
📦 Weekly Release 20260410
This release includes 67 commits. Key user-facing updates below.
New Features and Enhancements
- Introduced Prompt Rewrite & Translate feature for assisted input editing.
- Added Skill Panel with dedicated skills tab in the skill store and fixed skill icon rendering.
- Introduced lh notify CLI command for external agent callbacks.
- Added migrate openclaw CLI command.
- Added GraphAgent and agentFactory for graph-driven agent execution (experimental).
- New topic auto-creation every 4 hours for long-running sessions.
Models and Provider Expansion
- Added a new provider: StreamLake (快手万擎).
- Added GLM-5.1 model support with Kimi CodingPlan fixes.
- Added Seedance 2.0 & Seedance 2.0 Fast video generation models (pricing adjusted with 20% service fee).
- Expanded AIGC parameter support for image and video generation.
- Improved model type normalization for better provider compatibility.
- Multi-media and multiple connection mode support for ComfyUI integration.
Desktop Improvements
- Embedded CLI in the desktop app with PATH installation support.
- Added Electron version display in system tools settings.
- Fixed RuntimeConfig instant-apply working directory with recent list.
- Fixed desktop locale restore — now uses stored URL parameter instead of system locale.
- Improved remote re-auth for batched tRPC and clean OIDC on gateway disconnect.
Stability, Security, and UX Fixes
- Security: prevented path traversal in TempFileManager.writeTempFile; patched IDOR in addFilesToKnowledgeBase; upgraded better-auth with hardened humanIntervention requirement in builtin-tool-activator.
- Context engine: added typeof guard before .trim() calls to prevent runtime crashes.
- Agent runtime: preserved reasoning state across OpenAI providers; fixed service error serialization producing [object Object]; surfaced error reasonDetail in agent_runtime_end events.
- Knowledge Base: cleaned up vector storage when deleting knowledge bases.
- Templates: allow templates to specify policyLoad so default docs are fully injected.
- Skills: inject current agents information when lobehub_skill is activated; filter current agent out of available agents list; fix agents_documents overriding systemRole.
- Google Tools: use parametersJsonSchema for Google tool schemas.
- Web Crawler: prevent happy-dom CSS parsing crash in htmlToMarkdown.
- Mobile/UI: fixed video page icon collision, missing locale keys, model query param; hidden LocalFile actions on topic share page; allow manual close of hidden builtin tools.
- Auth: ENABLE_MOCK_DEV_USER now supported in checkAuth and openapi auth middleware.
- Sandbox: stopped using sanitizeHTMLContent to block scripts & sandbox styles.
Refactors
- Library/resource tree store for hierarchy and move sync.
- Removed legacy messageLoadingIds from chat store.
- Removed promptfoo configs and dependencies.
- OnboardingContextInjector wired into context engine.
Credits
Huge thanks to these contributors (alphabetical):
@arvinxx @canisminor1990 @cy948 @hardy-one @hezhijie0327 @Innei @MarcellGu @ONLY-yours @rdmclin2 @rivertwilight @sxjeru @tjx666

This release was automatically published from PR #13626.
Changes
See PR description: https://github.com/lobehub/lobehub/pull/13626
Commit Message
🚀 release: 20260407
This release includes 148 commits. Key updates are below.
- Response API tool execution is more capable and reliable — Added hosted builtin tools + client-side function tools and improved tool-call streaming/completion behavior. #13406 #13414 #13506 #13555
- Input and composition UX upgraded — Added AI input auto-completion and multiple chat-input stability fixes. #13458 #13551 #13481
- Model/provider compatibility improved — Better Gemini/Google tool schema handling and additional model updates. #13429 #13465 #13613
- Desktop and CLI reliability improved — Gateway WebSocket support and desktop runtime upgrades. #13608 #13550 #13557
- Security hardening continued — Fixed auth and sanitization risks and upgraded vulnerable dependencies. #13535 #13529 #13479
Models & Providers
- Added/updated support for glm-5v-turbo, GLM-5.1 updates, and qwen3.5-omni series. #13487 #13405 #13422
- Added additional ImageGen providers/models (Wanxiang 2.7 and Keling from Qwen). #13478
- Improved Gemini/Google tool schema and compatibility handling across runtime paths. #13429 #13465 #13613
Response API & Runtime
- Added hosted builtin tools in Response API and client-side function tool execution support. #13406 #13414
- Improved stream tool-call argument handling and response.completed output correctness. #13506 #13555
- Improved runtime error/context handling for intervention and provider edge cases. #13420 #13607
Desktop App
- Bumped desktop dependencies and runtime integrations (agent-browser, electron). #13550 #13557
- Simplified desktop release channel setup by removing nightly release flow. #13480
CLI
- Added OpenClaw migration command. #13566
- Added local device binding support for lh agent run. #13277
- Added WebSocket gateway support and reconnect reliability improvements. #13608 #13418
Security
- Removed risky apiKey fallback behavior in webapi auth path to prevent bypass risk. #13535
- Sanitized HTML artifact rendering and iframe sandboxing to reduce XSS-to-RCE risk. #13529
- Upgraded nodemailer to v8 to address SMTP command injection advisory. #13479
Bug Fixes
- Fixed image generation model default switch issues. #13587
- Fixed subtopic re-fork message scope behavior and agent panel reset edge cases. #13606 #13556
- Fixed chat-input freeze on paste and mention plugin behavior. #13551 #13415
- Fixed auth/social sign-in and settings UX edge cases. #13368 #13392 #13338
Credits
Huge thanks to these contributors:
@chriszf @hardy-one @Innei @LiJian @Neko @octopusnote @rdmclin2 @rivertwilight @RylanCai @suyua9 @sxjeru @Tsuki @WangYK @WindSpiritSR @Yizhuo @YuTengjing @hezhijie0327 @arvinxx

This release was automatically published from PR #13330.
Changes
See PR description: https://github.com/lobehub/lobehub/pull/13330
Commit Message
🚀 release: 20260326
This release includes 91 commits. Key updates are below.
- Agent can now execute background tasks — Agents can perform long-running operations without blocking your conversation. #13289
- Better error messages — Redesigned error UI across chat and image generation with clearer explanations and recovery options. #13302
- Smoother topic switching — No more full page reloads when switching topics while an agent is responding. #13309
- Faster image uploads — Large images are now automatically compressed to 1920px before upload, reducing wait times. #13224
- Improved knowledge base — Documents are now properly parsed before chunking, improving retrieval accuracy. #13221
Bot Platform
- WeChat Bot support — You can now connect LobeChat to WeChat, in addition to Discord. #13191
- Richer bot responses — Bots now support custom markdown rendering and context injection. #13294
- New bot commands — Added /new to start fresh conversations and /stop to halt generation. #13194
- Discord stability fixes — Fixed thread creation issues and Redis connection drops. #13228 #13205
Models & Providers
- GLM-5 is now available in the LobeHub model list. #13189
- Coding Plan providers — Added support for code planning assistant providers. #13203
- Tencent Hunyuan 3.0 ImageGen — New image generation model from Tencent. #13166
- Gemini content handling — Better handling when Gemini blocks content due to safety filters. #13270
- Claude token limits fixed — Corrected max window tokens for Anthropic Claude models. #13206
Skills & Tools
- Auto credential injection — Skills can now automatically request and use required credentials. #13124
- Smarter tool permissions — Built-in tools skip confirmation for safe paths like /tmp. #13232
- Model switcher improvements — Quick access to provider settings and visual highlight for default model. #13220
Memory
- Bulk delete memories — You can now delete all memory entries at once. #13161
- Per-agent memory control — Memory injection now respects individual agent settings. #13265
Desktop App
- Gateway connection — Desktop app can now connect to LobeHub Gateway for enhanced features. #13234
- Connection status indicator — See gateway connection status in the titlebar. #13260
- Settings persistence — Gateway toggle state now persists across app restarts. #13300
CLI
- API key authentication — CLI now supports API key auth for programmatic access. #13190
- Shell completion — Tab completion for bash/zsh/fish shells. #13164
- Man pages — Built-in manual pages for CLI commands. #13200
Security
- XSS protection — Sanitized search result image titles to prevent script injection. #13303
- Workflow hardening — Fixed potential shell injection in release automation. #13319
- Dependency update — Updated nodemailer to address security advisory. #13326
Bug Fixes
- Fixed skill page not redirecting correctly after import. #13255 #13261
- Fixed token counting in group chats. #13247
- Fixed editor not resetting when switching to empty pages. #13229
- Fixed manual tool toggle not working. #13218
- Fixed Search1API response parsing. #13207 #13208
- Fixed mobile topic menus rendering issues. #12477
- Fixed history count calculation for accurate context. #13051
- Added missing Turkish translations. #13196
Credits
Huge thanks to these contributors:
@bakiburakogun @hardy-one @Zhouguanyang @sxjeru @hezhijie0327 @arvinxx @cy948 @CanisMinor @Innei @LiJian @lobehubbot @Neko @rdmclin2 @rivertwilight @tjx666

This release was automatically published from PR #13295.
Changes
See PR description: https://github.com/lobehub/lobehub/pull/13295
Commit Message
This release includes a database schema migration involving 2 new tables for the Notification system.
Migration: Add Notification Tables
- Added 2 new tables: notifications, notification_deliveries
- Added notification jsonb column to user_settings
Notes for Self-hosted Users
- The migration runs automatically on application startup
- No manual intervention required
The migration owner: @tjx666 — responsible for this database schema change, reach out for any migration-related issues.

This release was automatically published from PR #13280.
Changes
See PR description: https://github.com/lobehub/lobehub/pull/13280
Commit Message
This release includes a database schema migration involving 6 new tables and 1 brief table for the Agent Task System.
Migration: Add Agent Task System Tables
- Added 6 new tables: tasks, task_dependencies, task_documents, task_topics, task_comments, briefs
- Added 24 foreign key constraints with proper cascade/set-null behaviors
- Added 35 indexes covering all major query patterns
- All sub-tables include user_id FK for row-level user isolation
Notes for Self-hosted Users
- The migration runs automatically on application startup
- No manual intervention required
The migration owner: @arvinxx — responsible for this database schema change, reach out for any migration-related issues.

This release was automatically published from PR #13155.
Changes
See PR description: https://github.com/lobehub/lobehub/pull/13155
Commit Message
This release includes ~90 commits. Key updates are below.
New Features and Enhancements
- Added Agent Documents implementation with knowledge management tools (#13093).
- Introduced context engine that injects referenced topic context into the last user message (#13104).
- Added server-side context compression for improved performance with long conversations (#12976).
- Introduced category-based mention menu in chat input for better discoverability (#13109).
- Added slash action tags, topic reference tool, and a command bus system (#12860).
- Added skill activity switching support (#12970).
- Improved agent context injection — skills discovery, device optimization, prompt cleanup (#13021).
Models and Providers
- Added XiaomiMiMo v2-pro and v2-omni support with web search enablement and tool merging (#13137, #13133, #13123).
- Added GPT-5.4 mini and nano models (#13094).
- Updated Grok 4.20 to 0309 and added MiniMax M2.7 (#13112).
- Added Seedance 1.5 Pro support (#13035).
- Updated Claude 4.6 series to 1M context window (#12994).
Desktop Improvements
- Added Linux window specialization (#13059).
- Display actual desktop app version with canary suffix (#13110).
- Removed electron-liquid-glass to fix click event blocking on macOS (#13070).
- Improved WelcomeStep layout centering in onboarding (#13125).
Performance
- Optimized search with BM25 indexes and ICU tokenizer (#12914).
- Optimized tool system prompt — removed duplicate APIs, simplified XML tags (#13041).
- Optimized user message renderer performance (#13108).
- Used incremental diff for snapshot messages to prevent OOM (#13136).
Image / Video
- Redesigned image / video generation UI (#13126).
- Added image/video mode switch (#13152).
Stability and Fixes
- Fixed duplicate AI provider ID: return CONFLICT instead of 500, added frontend form validation (#13153).
- Fixed builtin provider race condition with onConflictDoNothing on auto-init insert (#13153).
- Fixed "Last Used" badge showing on login page when only one login option exists (#13153).
- Fixed XiaomiMiMo chat payload — normalize reasoning content, filter empty user messages, drop unsupported trailing assistant messages (#13153).
- Fixed forkedFromIdentifier not stripped before LLM API request (#13142).
- Fixed empty editor state structure and wide screen layout (#13131).
- Fixed missing BusinessAuthProvider slot in auth layout (#13130).
- Fixed artifacts code scroll preservation while streaming (#13114).
- Fixed SSRF block error distinction from network errors (#13103).
- Fixed SPA HTML entry returning stale content after server upgrade (#12998).
- Fixed PDF.js worker loading from local assets via Vite ?url import (#13006).
- Fixed Responses API tool pairing and context limit errors (#13078).
- Fixed missing userId in embeddings API calls (#13077) and initModelRuntimeFromDB (#13100).
- Fixed internal thinking content filtering in openai-compatible payloads (#13067).
- Fixed unsupported xAI reasoning penalties pruning (#13066).
- Fixed input clearing immediately on send to preserve drafts during streaming (#13038).
- Fixed market OIDC lost call tools error (#13025).
- Fixed jsonb ? operator usage to avoid Neon rt_fetch bug (#13040).
- Fixed model provider popup problems (#13012).
- Fixed agent-level memory config priority over user settings (#13018).
- Fixed multi-provider model item selection (#12968).
- Fixed invokeBuiltinTool returning undefined instead of structured error (#13020).
- Fixed agent stream error in local dev (#13054).
- Fixed skill crash (#13011).
- Fixed desktop agent-browser upgrade to v0.20.1 (#12985).
- Fixed topic share modal inside router (#12951).
- Fixed Enter key submission during IME composition (#12963).
- Fixed error collapse default active key (#12967).
- Replaced per-message useNewScreen with centralized useConversationSpacer (#13042).
- Ref

This release was automatically published from PR #13032.
Changes
See PR description: https://github.com/lobehub/lobehub/pull/13032
Commit Message
This release includes a database schema migration adding BM25 full-text search indexes with ICU tokenizer across 14 tables.
Migration: Add BM25 Indexes with ICU Tokenizer
- Added BM25 indexes for 14 tables: agents, topics, files, knowledge_bases, user_memories, chat_groups, user_memories_contexts, user_memories_preferences, user_memories_activities, user_memories_identities, user_memories_experiences, user_memory_persona_documents, documents, messages
- All indexes use ICU tokenizer with English stemmer and stopwords for multilingual support
- Each index includes user_id as a keyword field with fast flag for efficient filter pushdown
- Enum/filter fields (type, status, memory_layer, memory_category, relationship, source_type, role) added as keyword+fast for filter pushdown into tantivy index scan
- chat_groups.content (system prompt) and messages.summary added as searchable text fields
- Large tables (documents, messages) are ordered last to avoid blocking smaller index builds
- All CREATE INDEX statements are preceded by DROP INDEX IF EXISTS for idempotency
Notes for Self-hosted Users
- The migration runs automatically on application startup
- Requires the pg_search extension (added in migration 0090)
- Index creation on large tables (messages, documents) may take several minutes depending on data volume
- No manual intervention required
The migration owner: @tjx666 — responsible for this database schema change, reach out for any migration-related issues.
Summary by Sourcery
Add a database migration that creates BM25 full-text search indexes with ICU tokenization across key content and metadata tables to improve search performance and filter pushdown.
New Features:
- Introduce BM25 full-text search indexes with ICU tokenizer on 14 user- and content-related tables for richer search capabilities.
- Make additional text fields such as chat group content and message summaries searchable via the new indexes.
Enhancements:
- Configure keyword+fast fields for user_id and other filterable enums to enable efficient filter pushdown during index scans.
- Order index creation so large tables are processed last and guard all index creation with DROP INDEX IF EXISTS for idempotent migrations.
Build:
- Add migration 0093 and its snapshot to the database migration system to apply the new BM25 indexes automatically on startup.

This release was automatically published from PR #12974.
Changes
See PR description: https://github.com/lobehub/lobehub/pull/12974
Commit Message
💻 Change Type
-
✨ feat
-
🐛 fix
-
♻️ refactor
-
💄 style
-
👷 build
-
⚡️ perf
-
✅ test
-
📝 docs
-
🔨 chore
🔗 Related Issue
N/A
🔀 Description of Change
- create stable*.yml from latest*.yml during stable desktop S3 publishing
- keep canary and nightly behavior unchanged
- preserve the existing version-prefixed manifest URLs used for archived assets
🧪 How to Test
-
Tested locally
-
Added/updated tests
-
No tests needed
Validated the release shell flow locally by simulating stable manifest generation and confirming stable.yml and stable-mac.yml are created with version-prefixed URLs.
📸 Screenshots / Videos
Before After
N/A N/A
📝 Additional Information
This is a hotfix for stable desktop release publishing to S3. Electron Builder emits latest*.yml for stable releases by default, while our update server expects stable*.yml under the stable/ channel path.
Summary by Sourcery
Ensure stable desktop releases published to S3 include correctly named update manifests for the stable channel.
Bug Fixes:
- Generate stable*.yml manifests from latest*.yml during stable channel S3 publishing so the update server can resolve stable update metadata correctly.
CI:
- Adjust desktop S3 publish GitHub Action to create stable*.yml manifests before URL rewriting and upload, keeping canary and nightly behavior unchanged.

This release was automatically published from PR #12956.
Changes
See PR description: https://github.com/lobehub/lobehub/pull/12956
Commit Message
This release includes ~400 commits. Key updates are below.
New Features and Enhancements
- Bot Platform Integration: Added abstract bot platform layer with QQ Bot, Telegram Bot, Lark/Feishu Bot, and Discord Bot integrations, including remote device support for IM integration.
- LobeHub CLI: Full CLI implementation across 5 phases — agent run/status, generate (text/image/video/TTS/ASR), doc, search, device, bot integration, cron, topic share, agent KB/file/pin, thread, and eval commands.
- Agent Skills: Added built-in skills management, skill store, agent browser automation skill, and tool detection.
- Video Generation: End-to-end video generation feature with free quota, webhook handling, and skeleton loading.
- Agent Benchmark: Added benchmark support with external scoring mode and dedicated DB schema.
- Memory Settings: Support for memory effort/tool permission configuration, user persona injection, and improved memory analysis.
- Batch Topic Deletion from file support.
- Runtime Config support for flexible deployment configuration.
- V1 API and Response API support (including OpenAI Responses API).
- Device Code Auth Flow for CLI authentication.
- Emoji Reactions for messages.
- Starter Suggested Questions and recommend agents.
- Page Tabs for Electron desktop.
- Sort Topics by Updated Time option.
- Change Email Address in profile settings.
- Model Detail Dropdown in model switch panel.
- Added unread completion indicator for agents and topics.
Models and Provider Expansion
- New providers: Straico, LongCat (美团).
- Added/updated model support:
- GPT-5.4 series
- Claude Sonnet 4.6 and Claude Opus 4.6 (including Bedrock)
- Gemini 3.1 Pro Preview and Gemini 3.1 Flash Lite Preview
- Qwen3.5 series (including Flash, OSS, and SiliconCloud models)
- Grok 4.20 series and Grok Imagine image generation
- Kimi K2.5 thinking models
- MiniMax 2.5 / MiniMax M2.5
- Nano Banana 2
- Seedream 5 Lite / Seedance 2.0
- NVIDIA new models
- GLM-5, GLM-4.6V, GLM-Image for Zhipu
- Additional Qwen image-to-image and text-to-image models
- Added video input support for SiliconCloud provider.
- Use Response API for Grok as default.
Desktop Improvements
- Integrated electron-liquid-glass for macOS Tahoe.
- Unified canary with stable app name/icon, added channel tag in About.
- Support clearing hotkey bindings in ShortcutManager.
- Subscription pages embedding with webview.
- Enhanced desktop menu and navigation system.
- Proactive token refresh on app startup and activation.
- DMG background image configuration.
- S3 publish for canary/nightly with cleanup.
- Unified update channel switching with S3 distribution.
Architecture and Infrastructure
- Vite SPA Migration: Migrated frontend from Next.js App Router to Vite SPA, restructured SPA routes to src/routes and src/router.
- Response API Support across agent runtime.
- Refactored client agent runtime and centralized tool availability checks.
- Added Redis pipeline support and Lua script execution.
- Database migrations: pg_search extension, video generation schema, agent skills schema, benchmark schema, topics description column, API key hash column, ID migration to nanoid.
- Preload bundled i18n resources with lazy-load for target language.
- Simplified build config, removed webpack customization, and resolved Vercel OOM.
- Class-based Zustand actions with flattenActions migration.
- Extracted @lobechat/local-file-shell shared package.
- Resolved all ESLint suppressions and enabled consistent-type-imports rule.
Stability, Security, and UX Fixes
- Fixed model provider popup problems and ModelSelect crash.
- Fixed tool engine, input-loading, and MCP tool install loading issues.
- Hardened Anthropic message building and sampling parameter handling.
- Fixed Vertex AI 400 error from duplicate tool function declarations.
- Fixed context window exceeded error dete

This release was automatically published from PR #12939.
Changes
See PR description: https://github.com/lobehub/lobehub/pull/12939
Commit Message
This release includes a database schema migration adding 1 new column to the topics table.
Migration: Add description column to topics table
- Added description (text, nullable) column to the topics table
- Uses idempotent ADD COLUMN IF NOT EXISTS syntax
Notes for Self-hosted Users
- The migration runs automatically on application startup
- No manual intervention required
The migration owner: @tjx666 — responsible for this database schema change, reach out for any migration-related issues.

This release was automatically published from PR #12862.
Changes
See PR description: https://github.com/lobehub/lobehub/pull/12862
Commit Message
This release includes a database schema migration adding a key_hash column to the api_keys table for authentication lookup.
Migration 0089: Add API Key Hash Column
- Added key_hash column (varchar(128), unique) to api_keys table
- Enables hash-based API key authentication lookup
Notes for Self-hosted Users
- The migration runs automatically on application startup
- No manual intervention required
- The migration is backwards-compatible — existing API keys will continue to work
The migration owner: @arvinxx — responsible for this database schema change, reach out for any migration-related issues.

This release was automatically published from PR #12757.
Changes
See PR description: https://github.com/lobehub/lobehub/pull/12757
Commit Message
This release includes 31 commits. Key updates are below.
New Features and Enhancements
- Added Telegram bot access support.
- Added electron page tabs functionality for desktop.
- Added device code auth flow for authentication.
- Added GPT-5.4 model support.
- Show last used auth provider on sign-in page for better UX.
- Support clearing hotkey bindings in desktop ShortcutManager.
- Added Gemini 3.1 Flash Lite Preview model and thinkingLevel5 extend param.
- Added auto aspect ratio and image search support for Nano Banana 2.
- User memories now default to inject user persona instead of identities.
Desktop Improvements
- Unified update channel switching with S3 distribution.
- Added S3 publish for canary/nightly and S3 cleanup (keep latest 15).
- Added electron page tabs functionality.
Stability and Fixes
- Fixed agents fork not working in community deploy.
- Fixed animation for single-line messages between reasoning and tool calls.
- Fixed Discord bot conflict with keyPrefix.
- Fixed skew plugin issue.
- Fixed userMemories database failure on extra structure mismatch.
- Fixed old LobeHub plugins update issue.
- Fixed context-engine tool type recovery from manifest when models strip suffixes.
- Added await to handleResponseAPIMode for proper error handling.
- Fixed M2M token for community agents/MCP/skill list.
- Fixed scripts to support Win32.
- Improved gateway and device gateway CI.
Credits
Huge thanks to these contributors (alphabetical):
@arvinxx @huangkairan @Innei @LiJian @Luis-Sambrano @nekomeowww @rdmclin2 @ReneWang @sxjeru @tjx666

This release was automatically published from PR #12755.
Changes
See PR description: https://github.com/lobehub/lobehub/pull/12755
Commit Message
💻 Change Type
-
✨ feat
-
🐛 fix
-
♻️ refactor
-
💄 style
-
👷 build
-
⚡️ perf
-
✅ test
-
📝 docs
-
🔨 chore
🔗 Related Issue
N/A
🔀 Description of Change
This hotfix updates the stable desktop release workflow on main to archive renderer assets from the current Electron renderer output directory instead of the removed out/ directory.
It prevents stable Linux desktop builds from failing after packaging succeeds when creating lobehub-renderer.tar.gz.
🧪 How to Test
-
Tested locally
-
Added/updated tests
-
No tests needed
Verified the workflow change by checking the final diff against main and running bunx prettier --check .github/workflows/release-desktop-stable.yml.
📸 Screenshots / Videos
Before After
N/A N/A
📝 Additional Information
This is a workflow-only hotfix for the stable release pipeline on main. The desktop packages were already building successfully; the failure came from the extra renderer archive step using an outdated path.
Made with Cursor
Summary by Sourcery
Build:
- Update Linux packaging step in the stable desktop release workflow to tar the renderer from apps/desktop/dist/renderer instead of the removed out directory, with a guard to ensure the directory exists.

This release was automatically published from PR #12532.
Changes
See PR description: https://github.com/lobehub/lobehub/pull/12532
Commit Message
💻 Change Type
-
✨ feat
-
🐛 fix
-
♻️ refactor
-
💄 style
-
👷 build
-
⚡️ perf
-
✅ test
-
📝 docs
-
🔨 chore
🔗 Related Issue
🔀 Description of Change
🧪 How to Test
-
Tested locally
-
Added/updated tests
-
No tests needed
📸 Screenshots / Videos
Before After
... ...
📝 Additional Information
Summary by Sourcery
Scope agent evaluation benchmarks and new bot provider bindings by user while introducing infrastructure for external chat platform integrations.
New Features:
- Add agent bot provider database schema and model to store per-agent external platform bindings with encrypted credentials.
- Introduce agent agency configuration types and persist agencyConfig on agents for Discord and Slack bot settings.
Bug Fixes:
- Associate agent evaluation benchmarks with a userId and restrict create, query, update, delete, and lookup operations to the owning user while still allowing access to system benchmarks.
Enhancements:
- Export new agentBotProvider schema and agency configuration types through existing database and type index barrels.
Tests:
- Extend agent evaluation benchmark model tests to cover user-scoped behavior, including isolation between different users and handling of system benchmarks.

This release includes 82 commits and 854 changed files. Key user-facing updates are below.
🚀 New Features and Enhancements
- Added Agent Benchmark support for more systematic agent performance evaluation.
- Introduced the video generation feature end-to-end, including entry points, sidebar “new” badge support, and skeleton loading for topic switching.
- Expanded memory capabilities: support for memory effort/tool permission configuration and improved timeout calculation for memory analysis tasks.
- Added desktop editor support for image upload via file picker.
🤖 Models and Provider Expansion
- Added a new provider: Straico.
- Added/updated support for:
- Claude Sonnet 4.6
- Gemini 3.1 Pro Preview
- Qwen3.5 series
- Grok Imagine (grok-imagine-image)
- MiniMax 2.5
- Added related i18n copy and model parameter adaptations.
🖥️ Desktop Improvements
- Integrated electron-liquid-glass (macOS Tahoe).
- Improved DMG background assets and desktop release workflow.
🛠️ Stability, Security, and UX Fixes
- Fixed multiple video generation pipeline issues: precharge refund handling, webhook token verification, pricing parameter usage, asset cleanup, and type safety.
- Fixed sanitizeFileName path traversal risks and added unit tests.
- Fixed MCP media URL generation with duplicated APP_URL prefix.
- Fixed Qwen3 embedding failures caused by batch-size limits.
- Fixed multiple UI/interaction issues, including mobile header agent selector/topic count, ChatInput scrolling behavior, and tooltip stacking context.
- Fixed missing @napi-rs/canvas native bindings in Docker standalone builds.
- Improved GitHub Copilot authentication retry behavior and response error handling in edge cases.
🙏 Thanks to Committers
Huge thanks to these contributors (alphabetical):
@AmAzing129 @Coooolfan @Innei @ONLY-yours @Zhouguanyang @arvinxx @eaten-cake @hezhijie0327 @nekomeowww @rdmclin2 @rivertwilight @sxjeru @tjx666

This release was automatically published from PR #12333.
Changes
See PR description: https://github.com/lobehub/lobehub/pull/12333
Commit Message
Add new models to NVIDIA NIM provider and simplify payload handling logic.
New Models:
- MiniMax-M2.1, DeepSeek V3.2, GLM-4.7, GLM-5, Kimi K2.5
Payload Handling:
- Remove redundant model list filtering
- Convert reasoning to reasoning_content (NVIDIA NIM API format requirement)
- Convert thinking.type to chat_template_kwargs.thinking (API parameter format)

This release was automatically published from PR #11708.
Changes
See PR description: https://github.com/lobehub/lobehub/pull/11708
Commit Message
💻 Change Type
-
✨ feat
-
🐛 fix
-
♻️ refactor
-
💄 style
-
👷 build
-
⚡️ perf
-
✅ test
-
📝 docs
-
🔨 chore
🔗 Related Issue
Closes #10346
🔀 Description of Change
1. 更新的文生图，图生图模型列表，z-image wan2.5 wan2.6 qwen-image-plus/max qwen-image-edit-plus/max
2. 新增 image2image endpoint，为老版本图生图模型进行兼容
3. 默认使用 multimodal-generation endpoint（新模型目前调研下来都是用这个了，同时支持图生图和文生图）
4. 支持多区域 Dashscope URL，跟随 baseUrl 参数，自动切分 /compatible-mode/v1 默认北京区域
北京 https://dashscope.aliyuncs.com
新加坡 https://dashscope-intl.aliyuncs.com
弗吉尼亚 https://dashscope-us.aliyuncs.com
Endpoint

multimodal-generation
text2image
🧪 How to Test
-
Tested locally
-
Added/updated tests
-
No tests needed
📸 Screenshots / Videos
Before After
... ...
📝 Additional Information
ref: https://help.aliyun.com/zh/model-studio/newly-released-models
ref: https://bailian.console.aliyun.com/cn-beijing/?tab=doc#/doc/?type=model&url=2987148
Summary by Sourcery
Extend Qwen image generation support to cover new text-to-image and image-to-image models while routing legacy models via dedicated text2image/image2image endpoints and defaulting other models to the multimodal-generation API.
New Features:
- Add model metadata and configuration for new Qwen image models including Z-Image Turbo, Qwen Image Edit Max/Plus, Qwen Image Max/Plus, and Wanxiang 2.5/2.6 variants.
- Introduce explicit handling of legacy text-to-image and image-to-image Qwen models via separate async text2image and image2image endpoints.
Enhancements:
- Update the Qwen image creation flow to prefer the multimodal-generation endpoint for newer models and improve error messaging and logging across image workflows.
- Reformat select Qwen chat model descriptions for consistency without changing behavior.
Tests:
- Adjust Qwen image creation tests to align with the new multimodal-generation behavior and removed strict input validation on qwen-image-edit-specific image URL requirements.

This release was automatically published from PR #12321.
Changes
See PR description: https://github.com/lobehub/lobehub/pull/12321
Commit Message
🩹 Hotfix v2.1.30
This PR starts a hotfix release from main.
Release Process
1. ✅ Hotfix branch created from main
2. ✅ Pushed to remote
3. 🔄 Waiting for PR review and merge
4. ⏳ Auto tag + GitHub Release will be created after merge

Created by hotfix script
Summary by Sourcery
Adjust release automation, linting configuration, and user panel UI behavior for hotfix v2.1.30.
Enhancements:
- Update UserPanel popover styling and behavior, including a skeleton loading state and adjusted placement and triggers.
- Refine PanelContent component typing and navigation event handling in the sidebar header layout.
- Extend lint-staged configuration to integrate eslint suppression pruning and ensure suppression files are committed for multiple file types.
CI:
- Simplify GitHub release creation to run for all tag types in the auto-tag workflow.
- Harden the desktop stable release workflow to safely handle missing release body content.

This hotfix was automatically published from PR #12260.
Changes
See PR description: https://github.com/lobehub/lobehub/pull/12260
Commit Message
🩹 Hotfix v2.1.28
This PR starts a hotfix release from main.
Release Process
1. ✅ Hotfix branch created from main
2. ✅ Pushed to remote
3. 🔄 Waiting for PR review and merge
4. ⏳ Auto tag + GitHub Release will be created after merge

Created by hotfix script
Summary by Sourcery
Introduce a dedicated hotfix workflow that automates creating hotfix branches, PRs, version bumps, and GitHub releases while aligning auto-tag and release processes with these hotfix semantics.
New Features:
- Add a hotfix workflow script and npm command to generate hotfix branches, push them, and open PRs against main with standardized titles and bodies.
- Extend the auto-tag workflow to detect hotfix PRs by branch naming, compute patch versions, update package.json, and create hotfix-specific tags and GitHub releases.
Enhancements:
- Simplify settings content staticization by always including business tab components instead of gating them by an environment flag.
- Adjust the release workflow to validate package.json version against the tag instead of mutating it during release, and update the release branch creation script to stop creating marker commits.
- Update dependency on @lobehub/ui to the latest minor version.
Build:
- Add a package.json script entry to run the new hotfix workflow CLI.
CI:
- Enhance the auto-tag GitHub Actions workflow to share tagging context via environment variables, tag the version-bump commit instead of the PR merge commit, and tailor release notes and tag messages for release vs hotfix PRs.

This release was automatically published from PR #12248.
Changes
See PR description: https://github.com/lobehub/lobehub/pull/12248
Commit Message
📦 Release v2.1.27
This branch contains changes for the upcoming v2.1.27 release.
Change Type
- Checked out from dev branch and merged to main branch
Release Process
1. ✅ Release branch created
2. ✅ Pushed to remote
3. 🔄 Waiting for PR review and merge
4. ⏳ Release workflow triggered after merge

Created by release script

Released on 2026-02-10
💄 Styles
- misc: Update i18n.

Released on 2026-02-09

Released on 2026-02-09
🐛 Bug Fixes
- misc: Fix multimodal content_part images rendered as base64 text.

Released on 2026-02-09
🐛 Bug Fixes
- swr: Prevent useActionSWR isValidating from getting stuck.
- misc: Fix editor content missing when send error, use custom avatar for group chat in sidebar.
💄 Styles
- misc: Update i18n.

Released on 2026-02-08
🐛 Bug Fixes
- misc: Register Notebook tool in server runtime.

Released on 2026-02-08
🐛 Bug Fixes
- misc: Add end-user info on OpenAI Responses API call, enable vertical scrolling for topic list on mobile.

Released on 2026-02-08
🐛 Bug Fixes
- misc: Add api/version and api/desktop to public routes, show notification when file upload fails due to storage plan limit.

Released on 2026-02-06
♻ Code Refactoring
- docker-compose: Restructure dev environment.
- misc: Upgrade agents/group detail pages tabs、hidden like button.
🐛 Bug Fixes
- misc: Fixed in community pluings tab the lobehub skills not display.
💄 Styles
- model-runtime: Add Claude Opus 4.6 support for Bedrock runtime.

Released on 2026-02-04
🐛 Bug Fixes
- model-runtime: Fix moonshot interleaved thinking and circular dependency.

Released on 2026-02-04
♻ Code Refactoring
- model-runtime: Extract Anthropic factory and convert Moonshot to RouterRuntime.

Released on 2026-02-04
🐛 Bug Fixes
- misc: Add the preview publish to market button preview check.

Released on 2026-02-04
🐛 Bug Fixes
- misc: Fix cannot uncompressed messages.

Released on 2026-02-03
🐛 Bug Fixes
- docker: Add librt.so.1 to fix PDF parsing.

Released on 2026-02-03
🐛 Bug Fixes
- changelog: Normalize versionRange to valid semver.

Released on 2026-01-31
🐛 Bug Fixes
- misc: Slove the group member agents cant set skills problem.

Released on 2026-01-31
🐛 Bug Fixes
- auth: Add AUTH_DISABLE_EMAIL_PASSWORD env to enable SSO-only mode.

Released on 2026-01-29
🐛 Bug Fixes
- model-bank: Fix ZenMux model IDs by adding provider prefixes.

Released on 2026-01-27
🐛 Bug Fixes
- misc: The klavis in onboarding connect timeout fixed.

We're thrilled to launch LobeHub - the ultimate space for work and life: to find, build, and collaborate with agent teammates that grow with you.
What new in the release?
- Multi-Agent Collaboration: Bring multiple specialized agents into one conversation. They debate, reason, and solve complex problems together—faster and smarter.
- Agent Builder: Describe what you want, and LobeHub builds the complete agent—skills, behavior, tools, and personality. No setup required.
- Pages: write, read and organize documents with Lobe AI
- Memory: Your agents remember your preferences, style, goals, and past projects—delivering uniquely personalized assistance that gets better over time.
- New Knowledge Base: Use folders to organize your knowledge & resource
- Marketplace: Publish, adopt, or remix agents in a thriving community where intelligence grows together.
- And even more

Released on 2025-12-04
🐛 Bug Fixes
- misc: Fix React CVE.

Released on 2025-10-30

Released on 2025-10-28
♻ Code Refactoring
- misc: Change files page from RSC to SPA mode to improve performance.
💄 Styles
- aihubmix: Update extendParams to include urlContext.
- misc: Update i18n.

Released on 2025-10-28

Released on 2025-10-27
💄 Styles
- misc: Add MiniMax-M2 model.

Released on 2025-10-27
💄 Styles
- misc: Adjust modal setting form styles for improved layout and responsiveness, Unzip file when uploading in knowledge base [LOB-500].

Released on 2025-10-26
💄 Styles
- misc: Improve provider modal height when creating custom provider.

Released on 2025-10-26
💄 Styles
- misc: Update i18n.

Released on 2025-10-24
✨ Features
- misc: Use env to control clerk allow origin feature.

Released on 2025-10-23
🐛 Bug Fixes
- misc: Loadmore not work & navbar not show in pwa.

Released on 2025-10-23
💄 Styles
- misc: Allow removal of top_p and similar request parameters.

Released on 2025-10-22

Released on 2025-10-22
♻ Code Refactoring
- misc: Fix model runtime cost calculate with CNY.

Released on 2025-10-21
💄 Styles
- settings: Broadcast locale changes and update switchLocale action.

Released on 2025-10-21
✨ Features
- misc: Add PDF export functionality to share modal.
🐛 Bug Fixes
- misc: Ignore abort signal errors in TRPC client, slove when pwa user info have code cannot be viewed in full.
💄 Styles
- misc: Add knowledge base mansory layout [LOB-496], improve rich text link display.

Released on 2025-10-21
🐛 Bug Fixes
- misc: Pass threadId to messages in sendMessageInServer.

Released on 2025-10-21
💄 Styles
- misc: Show message author in minimap.

Released on 2025-10-20
💄 Styles
- misc: Solve when desktop the sider agent list too long.

Released on 2025-10-20
♻ Code Refactoring
- i18n: Rm qa.
💄 Styles
- misc: Update i18n.

Released on 2025-10-19
✨ Features
- misc: Support image generation for siliconcloud.

Released on 2025-10-18
♻ Code Refactoring
- misc: Refactor upload router into lambda and decide to remove it in V2.

Released on 2025-10-18
🐛 Bug Fixes
- misc: Fix topic fetch not correct in custom agent.

Released on 2025-10-16
💄 Styles
- misc: Improve welcome message.

Released on 2025-10-16
🐛 Bug Fixes
- misc: Automatic topic creation switch does not work.

Released on 2025-10-16
✨ Features
- misc: Support Group Chat, Mention, and Multi-Agent Orchestration with feature flag.

Released on 2025-10-15
💄 Styles
- misc: Improve update notification.

Released on 2025-10-14
🐛 Bug Fixes
- misc: Update Claude workflows to use oauth token, vertext ai create image.

Released on 2025-10-14
💄 Styles
- misc: Add imagen model to vertex ai.

Released on 2025-10-14
🐛 Bug Fixes
- misc: Prevent Vertex AI JSON credentials from being split by comma.

Released on 2025-10-14
🐛 Bug Fixes
- misc: Fix mcp server connect issue and refactor web search implement, fix tools calling long name length >64 issue.

Released on 2025-10-12
✨ Features
- misc: Add new setting for default image num.

Released on 2025-10-12
🐛 Bug Fixes
- misc: Fix input cannot send markdown.
💄 Styles
- misc: Optimize OpenRouter modelFetch endpoint, update i18n.

Released on 2025-10-11
💄 Styles
- misc: Add more AWS regions, Update infini-ai models.

Released on 2025-10-11
💄 Styles
- misc: Add lab to support disable/enable rich text.

Released on 2025-10-10
🐛 Bug Fixes
- misc: Add 'gemini-2.5-flash-image' to disabled models Thinking.

Released on 2025-10-10
💄 Styles
- misc: Add delete & regenerate hotkeys.

Released on 2025-10-09

Released on 2025-10-09
✨ Features
- misc: Add new provider Cerebras.
🐛 Bug Fixes
- misc: Fix standalone plugin rerender issue.

Released on 2025-10-08
🐛 Bug Fixes
- desktop: Macos26 small icon.

Released on 2025-10-08
💄 Styles
- misc: Update i18n.

Released on 2025-10-07
♻ Code Refactoring
- misc: Refactor chat item.
💄 Styles
- misc: Add GPT-5 pro model.

Released on 2025-10-06
💄 Styles
- image: Optimize UX and fix fal pricing.

Released on 2025-10-06
🐛 Bug Fixes
- security: Sanitize Azure provider error responses to prevent API key exposure.
💄 Styles
- misc: Update i18n.

Released on 2025-10-05

Released on 2025-10-05
🐛 Bug Fixes
- misc: Type not preserved when model is sorted.

Released on 2025-10-05
💄 Styles
- misc: Update i18n.

Released on 2025-10-04
✨ Features
- misc: Support double-click to open multi agent window on the desktop.

Released on 2025-10-04
🐛 Bug Fixes
- misc: type not preserved when model is disabled or sorted.
💄 Styles
- misc: Nano banana support aspect_ratio.

Released on 2025-10-04
🐛 Bug Fixes
- misc: Custom provider fails when client requests are enabled.
💄 Styles
- misc: Optimized extendParams UI, update i18n.

Released on 2025-10-01
🐛 Bug Fixes
- misc: OllamaCloud error.
💄 Styles
- misc: Fix chat minimap overflow.

Released on 2025-09-30
💄 Styles
- misc: Add minimap to chat list for quick navigation.

Released on 2025-09-30
💄 Styles
- misc: Update i18n.

Released on 2025-09-29
✨ Features
- misc: Add builtin Python plugin, add Claude Sonnet 4.5 model to AI chat models.

Released on 2025-09-29

Released on 2025-09-27
🐛 Bug Fixes
- misc: Fix input empty group name.

Released on 2025-09-26
🐛 Bug Fixes
- misc: Resolve qwen-image-edit imageUrls conversion issue.

Released on 2025-09-25
🐛 Bug Fixes
- misc: Add proxyUrl configuration for NEW API provider.

Released on 2025-09-25

Released on 2025-09-24
💄 Styles
- misc: Enhanced Nvidia NIM chat experience, OpenAI models in AiHubMix use Responses API.

Released on 2025-09-24
🐛 Bug Fixes
- misc: Macos desktop sign.

Released on 2025-09-23

Released on 2025-09-22
♻ Code Refactoring
- misc: Move the ModelProvider to model-bank.

Released on 2025-09-22
💄 Styles
- misc: Enable thinkingBudget control for Vertex Gemini 2.5 models, update i18n.

Released on 2025-09-21
💄 Styles
- misc: Added AUTH_MICROSOFT_ENTRA_ID_BASE_URL routing.

Released on 2025-09-21
✨ Features
- misc: Support google video understanding.

Released on 2025-09-21
🐛 Bug Fixes
- misc: Update Responses search tool to web_search.

Released on 2025-09-19
✨ Features
- misc: Qwen provider add qwen-image-edit model support.

Released on 2025-09-18
🐛 Bug Fixes
- misc: Fix oidc open direct issue.

Released on 2025-09-18
✨ Features
- misc: Add scroll support for pinned assistants using ScrollShadow.

Released on 2025-09-18
🐛 Bug Fixes
- misc: Fix svg xss issue.

Released on 2025-09-16
♻ Code Refactoring
- misc: Improve db sql performance.
💄 Styles
- misc: Update SiliconCloud reasoning models.

Released on 2025-09-16
✨ Features
- misc: Support Vercel AI Gateway provider.

Released on 2025-09-15
💄 Styles
- misc: Improve error handle with agent config, support .doc file parse.

Released on 2025-09-15
💄 Styles
- misc: Enable toggling search on/off via search button click & historyCount button.

Released on 2025-09-14

Released on 2025-09-12
✨ Features
- misc: ChatInput support resize.

Released on 2025-09-11
🐛 Bug Fixes
- misc: Improve OpenAIStream processing to emit usage data for chunks lacking choices.

Released on 2025-09-11
♻ Code Refactoring
- misc: Refactor model runtime folder structure and add more tests.

Released on 2025-09-10
✨ Features
- misc: Seedream 4.0.
💄 Styles
- misc: Add hotkey tooltip to typobar actions.

Released on 2025-09-10
💄 Styles
- misc: Add CometAPI model provider and chat models, update i18n.

Released on 2025-09-09
🐛 Bug Fixes
- misc: Fix editor key handling.

Released on 2025-09-09
🐛 Bug Fixes
- misc: Fix Assistant List error message.

Released on 2025-09-08
✨ Features
- image: Implement model selection memory functionality.

Released on 2025-09-08
✨ Features
- misc: Add Math and TaskList to Editor.

Released on 2025-09-06
🐛 Bug Fixes
- misc: Revert V1 Mobile.

Released on 2025-09-06
♻ Code Refactoring
- misc: Refactor to remove edge runtime and add more tests.

Released on 2025-09-06
🐛 Bug Fixes
- misc: Enhance NewAPI with environment variables and fix routers compatibility.
💄 Styles
- misc: Update doubao-seed-1.6-vision models.

Released on 2025-09-05
♻ Code Refactoring
- misc: Remove edge runtime.

Released on 2025-09-05
🐛 Bug Fixes
- misc: Fix mobile header title to loog not ellipsis.

Released on 2025-09-05
💄 Styles
- misc: Update i18n.

Released on 2025-09-04
✨ Features
- misc: Add NewAPI as a router provider for multi-model aggregation.

Released on 2025-09-04
💄 Styles
- misc: Update i18n.

Released on 2025-09-04
✨ Features
- misc: Refactor to speed up send message in server mode.

Released on 2025-09-03
🐛 Bug Fixes
- misc: Fix socks5 proxy not work problem, fix virtuaso minheight was null.

Released on 2025-09-02
💄 Styles
- misc: Update i18n.

Released on 2025-09-01
💄 Styles
- misc: Add upload hint for non-visual model.

Released on 2025-09-01
🐛 Bug Fixes
- ai-image: Save config.imageUrl with fullUrl instead of key.

Released on 2025-09-01
💄 Styles
- misc: Adjust ControlsForm component to adapt to mobile phone display.

Released on 2025-08-31
♻ Code Refactoring
- misc: Remove base path.

Released on 2025-08-31
💄 Styles
- misc: Update i18n.

Released on 2025-08-30
♻ Code Refactoring
- misc: Remove webrtc sync feature flag.
✨ Features
- misc: Rename Gemini 2.5 flash image to Nano Banana.

Released on 2025-08-30
💄 Styles
- misc: Update DeepSeek V3.1 & Gemini 2.5 Flash Image Preview models.

Released on 2025-08-30
💄 Styles
- misc: Update i18n.

Released on 2025-08-29
🐛 Bug Fixes
- misc: Add Content-Security-Policy env.
💄 Styles
- misc: Support Gemini URL context tool.

Released on 2025-08-29
♻ Code Refactoring
- misc: Move chat item into chat.

Released on 2025-08-28
💄 Styles
- misc: Support html preview.

Released on 2025-08-28
🐛 Bug Fixes
- misc: Fix desktop route error.

Released on 2025-08-28

Released on 2025-08-27

Released on 2025-08-27
✨ Features
- misc: Add gemini 2.5 flash image for vertex ai.

Released on 2025-08-26
✨ Features
- image: Polish ai image.

Released on 2025-08-22
🐛 Bug Fixes
- files: Remove force-static rendering to enable session access.

Released on 2025-08-22
💄 Styles
- misc: Update mistral model vision ability.

Released on 2025-08-21

Released on 2025-08-21
♻ Code Refactoring
- misc: Move chain into @lobechat/prompts.

Released on 2025-08-19
✨ Features
- models: Add Qwen Image Edit model.

Released on 2025-08-19
🐛 Bug Fixes
- misc: Support Grok thinking models in AiHubMix, The 'stream_options' parameter is only allowed when 'stream' is enabled.

Released on 2025-08-18
🐛 Bug Fixes
- mcp: Use customParams for environment settings fallback.

Released on 2025-08-17
🐛 Bug Fixes
- db: Desktop local db can't vectorization.

Released on 2025-08-17
✨ Features
- provider: Add BFL provider support for image generation.

Released on 2025-08-16
♻ Code Refactoring
- misc: Refactor prompts folder to the @lobechat/prompts pacakge.

Released on 2025-08-16
♻ Code Refactoring
- misc: Refactor const folder to a new package.

Released on 2025-08-15
✨ Features
- feature-flags: Add ai_image flag to control AI painting UI.

Released on 2025-08-14

Released on 2025-08-13
💄 Styles
- misc: Update Mistral AI models & Optimize many model providers fetching.

Released on 2025-08-12
💄 Styles
- misc: Adjust near bottom size on thinking scroll.

Released on 2025-08-12
💄 Styles
- misc: Improve Gemini error display with promptFeedback.

Released on 2025-08-11
💄 Styles
- misc: Support new GPT-5 Verbosity params.

Released on 2025-08-11
🐛 Bug Fixes
- misc: Solve the cache problem caused by the same dom id when sharing pictures.

Released on 2025-08-10
💄 Styles
- misc: Add mask effect to thinking scroll, update i18n.

Released on 2025-08-09
🐛 Bug Fixes
- pricing: Adjust cachedInput values for GPT-5 models.

Released on 2025-08-08
♻ Code Refactoring
- pricing: Introduce new pricing system.

Released on 2025-08-08
✨ Features
- misc: Add GPT-5 series models.

Released on 2025-08-07
🐛 Bug Fixes
- misc: Missing languages it-IT, pl-PL, nl-NL.

Released on 2025-08-07
🐛 Bug Fixes
- misc: Optimize Gemini error message display & Filter empty messages.

Released on 2025-08-06
♻ Code Refactoring
- misc: Refactor trace type.

Released on 2025-08-06
🐛 Bug Fixes
- misc: Fix ollama model output without thinking.
💄 Styles
- misc: Add Claude Opus 4.1 model, update i18n.

Released on 2025-08-05
✨ Features
- misc: Support gpt-oss in ollama provider.

Released on 2025-08-05
🐛 Bug Fixes
- misc: Provider config checker uses outdated API key.

Released on 2025-08-05
🐛 Bug Fixes
- misc: Fix remote avatar broken in desktop.
💄 Styles
- misc: Update mask style.

Released on 2025-08-05
🐛 Bug Fixes
- misc: Break line for Gemini Artifacts.

Released on 2025-08-04
💄 Styles
- misc: Update models.

Released on 2025-08-03
🐛 Bug Fixes
- misc: Aihubmix provider request headers.

Released on 2025-08-02
♻ Code Refactoring
- misc: Move types to separate package.
🐛 Bug Fixes
- desktop: Settings window can't exit when fullscreen.

Released on 2025-08-01
💄 Styles
- misc: Update i18n.

Released on 2025-07-30
💄 Styles
- misc: Improve mcp plugin calling and display.

Released on 2025-07-29
🐛 Bug Fixes
- misc: Moonshot assistant messages must not be empty.
💄 Styles
- misc: Add volcengine kimi-k2 model, Add Zhipu GLM-4.5 models.

Released on 2025-07-29
♻ Code Refactoring
- misc: Clean mcp sitemap, refactor jose-JWT to xor obfuscation.
💄 Styles
- misc: Add more OpenAI SDK Text2Image providers, update i18n.

Released on 2025-07-29
💄 Styles
- misc: Support more Text2Image from Qwen.

Released on 2025-07-28
✨ Features
- misc: Implement API Key management functionality.

Released on 2025-07-26
💄 Styles
- misc: Add Gemini 2.5 Flash-Lite GA model.

Released on 2025-07-25
🐛 Bug Fixes
- misc: Update convertUsage to handle XAI provider and adjust OpenAIStream to pass provider.

Released on 2025-07-24
✨ Features
- misc: Support custom hotkey on desktop.

Released on 2025-07-24
🐛 Bug Fixes
- misc: Fix chat stream in desktop and update shortcut.
💄 Styles
- misc: Add cached token count to usage of GoogleAI and VertexAI, fix desktop titlebar style in window, fix sub topic width in md responsive.

Released on 2025-07-23
💄 Styles
- misc: Update i18n.

Released on 2025-07-22
✨ Features
- misc: Add Qwen image generation capabilities.

Released on 2025-07-22
♻ Code Refactoring
- misc: Add badge and improve document.
💄 Styles
- misc: Update tray icon.

Released on 2025-07-22
🐛 Bug Fixes
- misc: Remove debug logging from ModelRuntime and async caller.

Released on 2025-07-21
🐛 Bug Fixes
- groq: Enable streaming for tool calls and add Kimi K2 model.
💄 Styles
- misc: Modal list header sticky style.

Released on 2025-07-21
💄 Styles
- misc: Fix lobehub provider /chat in desktop.

Released on 2025-07-19
🐛 Bug Fixes
- misc: Try fix authorization code exchange & pin next-auto to beta.29.

Released on 2025-07-19
✨ Features
- misc: Add zhipu cogview4.
🐛 Bug Fixes
- misc: Some ai image bugs.

Released on 2025-07-17
🐛 Bug Fixes
- misc: Use server env config image models.

Released on 2025-07-17
✨ Features
- misc: Refactor desktop oauth and use JWTs token to support remote chat.
Breaking: server instance is need to upgrade to >=1.100.0 , or it will break

Released on 2025-07-15
🐛 Bug Fixes
- misc: Some ai image generation feedback issues.

Released on 2025-07-14
✨ Features
- plugin: Support Streamable HTTP MCP Server Auth.
- misc: support AI Image.

Released on 2025-07-14
💄 Styles
- misc: Update i18n.

Released on 2025-07-14
💄 Styles
- misc: Fix discover translation.

Released on 2025-07-13
✨ Features
- misc: Add network proxy for desktop.

Released on 2025-07-12
💄 Styles
- misc: Support new Doubao thinking models, update i18n.

Released on 2025-07-11
🐛 Bug Fixes
- misc: Grok-4 reasoning model universal matching.

Released on 2025-07-11
💄 Styles
- misc: Open new topic by tap Just Chat again.

Released on 2025-07-10
💄 Styles
- misc: Integrate Amazon Cognito for user authentication.

Released on 2025-07-10
💄 Styles
- misc: Add grok-4-0709 model from xAI, fix theme issue in desktop.

Released on 2025-07-10

Released on 2025-07-10
🐛 Bug Fixes
- misc: Fix locale hydration error in SSR.

Released on 2025-07-08
✨ Features
- misc: Add MCP marketplace and mcp plugin one-click installation in desktop.

Released on 2025-07-02
🐛 Bug Fixes
- misc: Wrong Gemini 2.5 Pro thinkbudget.

Released on 2025-07-01
♻ Code Refactoring
- misc: Migrate to @google/genai SDK for Google Gemini API and Vertex AI.

Released on 2025-07-01
💄 Styles
- misc: Update i18n.

Released on 2025-06-30
🐛 Bug Fixes
- misc: Pin antd@5.26.2 to fix build error.
💄 Styles
- misc: Add DeepResearch models from OpenAI.

Released on 2025-06-28
🐛 Bug Fixes
- mermaid: Firefox mermaid show error.

Released on 2025-06-28
🐛 Bug Fixes
- misc: Fix desktop chunk issue.

Released on 2025-06-23
🐛 Bug Fixes
- misc: Google Gemini tools declarations.

Released on 2025-06-23
💄 Styles
- misc: Optimized Gemini thinkingBudget configuration.

Released on 2025-06-23
💄 Styles
- misc: Add blockAds & stealth params for Browserless.

Released on 2025-06-23
💄 Styles
- misc: Update i18n.

Released on 2025-06-22
🐛 Bug Fixes
- misc: Remove unsupported parameters of Hunyuan.

Released on 2025-06-22

Released on 2025-06-22
💄 Styles
- openrouter: Add stable versions of Gemini 2.5 models.
- misc: Update i18n.

Released on 2025-06-21
🐛 Bug Fixes
- misc: Fix MiniMax-M1 reasoning tag missing.

Released on 2025-06-20
💄 Styles
- misc: Update i18n.

Released on 2025-06-19
🐛 Bug Fixes
- misc: Correctly pass reasoning.summary.

Released on 2025-06-19
💄 Styles
- misc: Update model card for Gemini 2.5 Pro via OpenRouter.

Released on 2025-06-19
💄 Styles
- misc: Update i18n.

Released on 2025-06-18
💄 Styles
- misc: Update Gemini 2.5 Pro, Flash GA models. Add Gemini 2.5 Flash-Lite Preview model.

Released on 2025-06-17
🐛 Bug Fixes
- misc: Enhance the multi-display window opening experience.

Released on 2025-06-15
🐛 Bug Fixes
- misc: Improve chat selectors and enhance topic handling logic.

Released on 2025-06-15
🐛 Bug Fixes
- misc: Correctly handle reasoning_effort.

Released on 2025-06-12
💄 Styles
- misc: Add Doubao Seed 1.6 model.

Released on 2025-06-12
🐛 Bug Fixes
- chat: Improve response animation merging logic.
💄 Styles
- misc: Support web_search_preview & fix some bug form OpenAI Response API.

Released on 2025-06-11
💄 Styles
- misc: Transition animation switch.

Released on 2025-06-11

Released on 2025-06-10
💄 Styles
- ModelSelect: Improve mobile layout and text overflow handling.
- misc: Update pplx abilities tags, support vision.

Released on 2025-06-10
♻ Code Refactoring
- misc: Refactor branding info.

Released on 2025-06-09
♻ Code Refactoring
- misc: Refactor <think> & </think> handling.

Released on 2025-06-08
🐛 Bug Fixes
- misc: Fix openai default Responses API issue.

Released on 2025-06-07
💄 Styles
- misc: Add support to azureopenai embedding.

💄 Styles
- misc: Add Volcengine & OpenAI-like Provider (e.g. oneapi) model fetch support.