Lemonade Server: download ældre versioner

Headline
- TheNoise image backend promoted to Core tier for ROCm on AMD GPUs across Linux and Windows.
- Image editing now accepts multiple reference images with automatic size selection and lossless passthrough.
- Live inference metrics added to GET /v1/stats, showing active requests and estimated tokens per second.
- AMD GPU utilization and VRAM telemetry now reported on Windows alongside NVIDIA in system-stats.
Breaking Changes
- GET /api/v1/system-stats now reports gpu_percent and vram_gb for the primary GPU by total memory instead of the GPU with the highest utilization — on multi-GPU systems these fields will show different values, requiring updates to dashboards and alerting rules.
- Models gated behind a guest-tier backend are now hidden from GET /v1/models?show_all=true and return model_not_supported on pull until the guest backend is installed — existing setups relying on those models being visible will no longer see them.
What's Changed
Thanks @bitgamma, @fl0rianr, @jeremyfowers, @joshuacox, @sreeram-11, @zaneni6 for your awesome contributions to this release!
Click to expand changelog
- refactor(server): run backend servers through ServerProcess and NativeProcess by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/3705
- fix(server): support multi-reference image editing by @fl0rianr in https://github.com/lemonade-sdk/lemonade/pull/3801
- fix(config): read the distro defaults override from lemond's own install prefix by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/3775
- feat(images): add per-model edit size caps and Qwen-Image-2.1 Turbo INT8 by @fl0rianr in https://github.com/lemonade-sdk/lemonade/pull/3803
- fix(models): preserve catalog identities in extra_models_dir by @fl0rianr in https://github.com/lemonade-sdk/lemonade/pull/3805
- fix(models): prevent duplicate discovery of incomplete shards by @fl0rianr in https://github.com/lemonade-sdk/lemonade/pull/3811
- docs: add systemd start step to the Arch install page by @joshuacox in https://github.com/lemonade-sdk/lemonade/pull/3784
- ci: bump auto-labeler to Claude Haiku 5.5 by @fl0rianr in https://github.com/lemonade-sdk/lemonade/pull/3774
- feat(backends): label each backend with a tier and a format by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/3706
- Update Lemonade to ROCm 10.1.0 by @sreeram-11 in https://github.com/lemonade-sdk/lemonade/pull/3821
- chore(flm): bump FLM to v1.0.8 by @zaneni6 in https://github.com/lemonade-sdk/lemonade/pull/3810
- Build cpp-httplib once instead of in every file that includes it by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/3837
- fix(server): report NVIDIA GPU and VRAM usage by @fl0rianr in https://github.com/lemonade-sdk/lemonade/pull/3282
- mark TheNoise as Core-tier by @bitgamma in https://github.com/lemonade-sdk/lemonade/pull/3826
- test: pre-pull server_llm.py's models and cover implicit pull on inference by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/3839
- lemonade bench: add prefill TPS by @bitgamma in https://github.com/lemonade-sdk/lemonade/pull/3759
- Add AnythingLLM and Phoronix Lemonade 11.9 articles to news page by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/3836
- fix(server): expose live inference throughput across backends by @fl0rianr in https://github.com/lemonade-sdk/lemonade/pull/3791
- test: isolate every C++ test from the real config dir by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/3612
New Contributors
- @joshuacox made their first contribution in https://github.com/lemonade-sdk/lemonade/pull/3784
Full Changelog: https://github.com/lemonade-sdk/lemonade/compare/v2026.42.1...v2026.43.3

Windows installers are signed. Free code signing provided by SignPath.io, certificate by SignPath Foundation. See our Code Signing Policy.

Headline
- Streaming requests no longer hang when a client disconnects mid-request.
- Windows support added for TheNoise (ROCm) via a bundled portable CPython interpreter.
- extra_models_dir now correctly import models from Hugging Face style cache directories.
- Qwen3.8-27B support for FLM
Breaking Changes
- The FastFlowLM NPU model hy-mt2:1.8b was renamed to hy-mt2-flash:1.8b; update any configs or scripts that reference the old tag.
What's Changed
Thanks @AaronStGeorge, @bitgamma, @github-actions, @jeremyfowers, @julianxhokaxhiu, @ramkrishna2910 for your awesome contributions to this release!
Click to expand changelog
- Auto-install specified backends with lemonade bench by @bitgamma in https://github.com/lemonade-sdk/lemonade/pull/3690
- test: remove flaky test_000b_retired_endpoints_removed by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/3720
- Add Windows support for TheNoise by @bitgamma in https://github.com/lemonade-sdk/lemonade/pull/3666
- Don't fail HTTPS requests on Windows when cert revocation status is unknown by @ramkrishna2910 in https://github.com/lemonade-sdk/lemonade/pull/3716
- fix(models): group shard files at the extra_models_dir root by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/3656
- fix(server): watch extra_models_dir's nested folders on Linux and macOS by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/3726
- docs(models): specify extra_models_dir behavior in the custom model guide by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/3633
- test(models): describe extra_models_dir layouts declaratively by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/3634
- feat(hrx): bump hrx to b99 + add 26 models by @AaronStGeorge in https://github.com/lemonade-sdk/lemonade/pull/3723
- chore(backends): bump FastFlowLM to v1.0.7 by @julianxhokaxhiu in https://github.com/lemonade-sdk/lemonade/pull/3744
- fix(http): make download stall timeouts follow global_timeout by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/3730
- fix(models): give a shared folder name to one model in both forms by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/3657
- Change evaluation date for AI content policy by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/3760
- bump TheNoise version by @bitgamma in https://github.com/lemonade-sdk/lemonade/pull/3750
- feat(models): support Hugging Face and ModelScope caches in extra_models_dir by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/3635
- fix(tray): make --spawn-server a no-op on macOS by @ramkrishna2910 in https://github.com/lemonade-sdk/lemonade/pull/3756
- docs: add the container backends spec by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/3697
- Update llama.cpp to b11393 by @github-actions[bot] in https://github.com/lemonade-sdk/lemonade/pull/3754
Full Changelog: https://github.com/lemonade-sdk/lemonade/compare/v2026.41.1...v2026.42.1

Windows installers are signed. Free code signing provided by SignPath.io, certificate by SignPath Foundation. See our Code Signing Policy.

Headline
- ROCm backends move to ROCm 10 for llama.cpp, whisper.cpp, stable-diffusion.cpp, ThinkSound, TRELLIS, and ACE-Step.
- Non-streaming requests now stop as soon as the client disconnects, so an abandoned request no longer ties up the server.
- Kokoro TTS fixes four British English voices and one French voice that returned short fixed-length audio.
- Configuration now rejects unknown backend keys instead of silently ignoring them.
Breaking Changes
- ROCm backends now use ROCm 10. The first ROCm model load after upgrading downloads the new runtime (about 2.6 GB).
- lemonade config set now rejects unknown backend keys, for example flm.flm_bin.
What's Changed
Thanks @RaulMermans, @abn, @bitgamma, @jeremyfowers, @pwilkin, @ramkrishna2910, @superm1 for your awesome contributions to this release!
Click to expand changelog
- feat(tray): add --spawn-server to natively supervise a local lemond by @abn in https://github.com/lemonade-sdk/lemonade/pull/3304
- Fix repository documentation links and published docs paths by @ramkrishna2910 in https://github.com/lemonade-sdk/lemonade/pull/3665
- Default release.py to unsigned tags by @superm1 with @Copilot in https://github.com/lemonade-sdk/lemonade/pull/3660
- Move stable diffusion back to our fork and move to ROCm 10 by @superm1 in https://github.com/lemonade-sdk/lemonade/pull/3649
- docs: link website-only pages to the website, revert marketplace link rewriting by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/3667
- Filter qwen3.6-moe-35b-a3b-FLM on Windows systems under 64 GB RAM by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/3675
- fix(server): propagate non-streaming request cancellation to backends by @abn in https://github.com/lemonade-sdk/lemonade/pull/2898
- Update REPO_MANAGER_VERSION to v1.0.4 by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/3692
- fix: change deprecated --no-mmap option to --load-mode none for llama.cpp by @pwilkin in https://github.com/lemonade-sdk/lemonade/pull/3559
- fix(config): reject invalid backend-specific keys by @RaulMermans in https://github.com/lemonade-sdk/lemonade/pull/3684
- bump Kokoro by @bitgamma in https://github.com/lemonade-sdk/lemonade/pull/3695
- docs: add a spec writing guide under docs/dev/specs by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/3687
New Contributors
- @RaulMermans made their first contribution in https://github.com/lemonade-sdk/lemonade/pull/3684

Windows installers are signed. Free code signing provided by SignPath.io, certificate by SignPath Foundation. See our Code Signing Policy.

Headline
- Streaming models now run on AMD integrated GPUs by sizing against the APU GTT pool instead of the fixed VRAM carve-out, enabling model streaming on hardware previously unsupported.
- Junie by JetBrains is a new launch agent available via lemonade launch junie, with profile generation and integration documentation.
- Cloud provider context window and completion token limits are now parsed from model metadata and surfaced on GET /v1/models.
- AMD GPU entries in GET /api/v1/system-info show a marketing name (e.g. "AMD Radeon RX 9070 XT (gfx1201)") instead of a raw numeric code.
Breaking Changes
- 'openmoss:rocm' and sd-cpp:rocm temporarily removed while we work on a couple of bugs.
What's Changed
Thanks @abn, @apollo-2006, @atirna, @bong-water-water-bong, @fl0rianr, @jeremyfowers, @mashan555, @noamsto, @superm1, @zaneni6 for your awesome contributions to this release!
Click to expand changelog
- docs(embeddable): update packaging-time bundling limitations by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/3548
- Report the AMD GPU marketing name on Linux by @apollo-2006 in https://github.com/lemonade-sdk/lemonade/pull/3601
- fix(server): size streaming models against APU GTT pool, not VRAM carve-out (#3377) by @bong-water-water-bong in https://github.com/lemonade-sdk/lemonade/pull/3502
- fix: dispatch image upscaling through backend capabilities by @atirna in https://github.com/lemonade-sdk/lemonade/pull/3511
- feat(models): add Ornith 1.5 9B and 35B A3B to the catalogue by @noamsto in https://github.com/lemonade-sdk/lemonade/pull/3564
- Don't run inference backend jobs on release branch builds by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/3604
- fix(server): drop the lookup-miss registry reload by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/3613
- Remove OpenMOSS ROCm support on all operating systems by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/3615
- feat(server): parse and propagate cloud model token limits by @abn in https://github.com/lemonade-sdk/lemonade/pull/3568
- refactor(mcp): make tool definitions declarative by @fl0rianr in https://github.com/lemonade-sdk/lemonade/pull/3619
- fix: support mbedTLS 4 on Arch Linux by @fl0rianr in https://github.com/lemonade-sdk/lemonade/pull/3188
- chore(backends): bump FastFlowLM to v1.0.6 by @zaneni6 in https://github.com/lemonade-sdk/lemonade/pull/3623
- Add Junie by JetBrains support to lemonade launch by @mashan555 in https://github.com/lemonade-sdk/lemonade/pull/3621
- Remove sd-cpp ROCm support on Windows by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/3616
- docs: add stable platform install pages by @fl0rianr in https://github.com/lemonade-sdk/lemonade/pull/2923
- ci(docs): check heading anchors by replacing markdown-link-check with lychee by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/3632
- docs(models): name the custom model guide's sections after their topics by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/3640
- ci: test PPA package upgrade path (config + model persistence) by @superm1 in https://github.com/lemonade-sdk/lemonade/pull/3396
New Contributors
- @atirna made their first contribution in https://github.com/lemonade-sdk/lemonade/pull/3511
- @mashan555 made their first contribution in https://github.com/lemonade-sdk/lemonade/pull/3621
Full Changelog: https://github.com/lemonade-sdk/lemonade/compare/v2026.39.1...v2026.40.0

Windows installers are signed. Free code signing provided by SignPath.io, certificate by SignPath Foundation. See our Code Signing Policy.

Headline
- TheNoise updated to 0.7.1 with support for image editing and additional AMD GPUs: gfx103X, gfx110X, and gfx120X.
- allowed_origins now accepts :* wildcard ports and *.domain subdomain patterns for local development and homelab setups.
- New /v1/docs endpoint serves browsable API reference documentation with a JSON index and individual markdown pages.
- Configurable VRAM auto-eviction via auto_evict and auto_evict_threshold_pct, with prompt cache now preserved during soft-idle transitions.
- Prerelease candidate builds available for early testing via GitHub prereleases under candidate-v<version> tags and the Docker candidate image tag.
Breaking Changes
- Version format changed from X.Y.Z to YYYY.WW.N or YYYY.WW.0~<count>.<hash>; tools and scripts that parse --version output or compare version strings must be updated.
Lemonade Server

| Operating System | Downloads |
| ------------------ | -------------------------------------------- |
| Windows | lemonade.msi |
| Ubuntu 24.04+ | Launchpad PPA |
| Debian 13 (x86_64) | lemonade-server_2026.39.1-debian13_amd64.deb |
| Debian 13 (ARM64) | lemonade-server_2026.39.1-debian13_arm64.deb |
| Fedora 43 (x86_64) | lemonade-server-2026.39.1-fc43.x86_64.rpm |
| Fedora 43 (ARM64) | lemonade-server-2026.39.1-fc43.aarch64.rpm |
| Fedora 44 (x86_64) | lemonade-server-2026.39.1-fc44.x86_64.rpm |
| Fedora 44 (ARM64) | lemonade-server-2026.39.1-fc44.aarch64.rpm |
| macOS | Lemonade-2026.39.1-Darwin.pkg |

Other platforms? See our Installation Options for Docker, Snap, Arch, Debian, and more.
Embeddable Lemonade
Portable binaries for bundling into your own installer. Run lemond ./ as a subprocess.

| Platform | Download |
| ------------ | ------------------------------------------------- |
| Ubuntu x64 | lemonade-embeddable-2026.39.1-ubuntu-x64.tar.gz |
| Ubuntu arm64 | lemonade-embeddable-2026.39.1-ubuntu-arm64.tar.gz |
| Windows x64 | lemonade-embeddable-2026.39.1-windows-x64.zip |
| macOS arm64 | lemonade-embeddable-2026.39.1-macos-arm64.tar.gz |

What's Changed
Thanks @AaronStGeorge, @Bekhouche, @GabrielReusRodriguez, @SlawomirNowaczyk, @Yigtwxx, @abn, @anditherobot, @antmikinka, @bitgamma, @fl0rianr, @github-actions, @iswaryaalex, @jamespthomas, @jeremyfowers, @kenvandine, @meghsat, @noamsto, @popey, @sofiageo, @superm1, @wariobot09, @zaneni6 for your awesome contributions to this release!
Click to expand changelog
- fix(server): streaming stall window follows global_timeout (#3386) by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/3477
- ci(hrx): bump hrx + add llama-3.1 8b ci coverage by @AaronStGeorge in https://github.com/lemonade-sdk/lemonade/pull/3503
- feat: update FLM to v1.0.4 by @zaneni6 in https://github.com/lemonade-sdk/lemonade/pull/3482
- Switch to bundler module resolution for tsconfig by @superm1 in https://github.com/lemonade-sdk/lemonade/pull/3186
- server(mcp): revert internal mcp client code (from PR 2605) by @fl0rianr in https://github.com/lemonade-sdk/lemonade/pull/3498
- Minor follow-up cleanup to #3330 by @SlawomirNowaczyk in https://github.com/lemonade-sdk/lemonade/pull/3444
- fix(routing): don't drop the whole router policy over one hardware-filtered classifier by @Bekhouche in https://github.com/lemonade-sdk/lemonade/pull/3414
- ci(vllm): keep the HF model cache outside the workspace by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/3345
- update image endpoints documentation by @bitgamma in https://github.com/lemonade-sdk/lemonade/pull/3358
- fix: name the gpt-oss-120b checkpoint instead of globbing by @noamsto in https://github.com/lemonade-sdk/lemonade/pull/3496
- feat(server): support wildcard port and host matching in allowed origins by @abn in https://github.com/lemonade-sdk/lemonade/pull/3470
- Remove the Linux Distro Builds ci job by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/3524
- Use the selected GPU for context auto-tuning by @popey in https://github.com/lemonade-sdk/lemonade/pull/3018
- Fix i

Headline
- This quality release fixes two top bugs: allowed_origins no longer blocks remote use, and the fixed 120s prefill timeout is now a 600s default that is configurable via global_timeout.
- A new experimental llamacpp-hrx backend runs AMD's HRX runtime in early access on AMD gfx1100/gfx1151 GPUs on Linux.
- The pinned llama.cpp backend version was updated to add Qwen3-Next support.
- Models placed in chat/, embeddings/, and reranking/ subfolders of extra_models_dir are now classified by type and listed individually.
Breaking Changes
- extra_models_dir models in reserved folders (chat/, embeddings/, reranking/) are now listed individually per file instead of as a single extra.<directory> model; legacy extra.chat/extra.embeddings/extra.reranking IDs still resolve as hidden aliases to the first file alphabetically.
- The LEMONADE_ALLOWED_ORIGINS environment variable is deprecated in favor of the allowed_origins config/CLI option and auto-migrates at startup with a warning; migrate to the new option, as the environment variable is only supported temporarily.
Lemonade Server

| Operating System | Downloads |
| ------------------ | ----------------------------------------- |
| Windows | lemonade.msi |
| Ubuntu 24.04+ | Launchpad PPA |
| Debian 13 (x86_64) | lemonade-server_11.9.0-debian13_amd64.deb |
| Debian 13 (ARM64) | lemonade-server_11.9.0-debian13_arm64.deb |
| Fedora 43 (x86_64) | lemonade-server-11.9.0-fc43.x86_64.rpm |
| Fedora 43 (ARM64) | lemonade-server-11.9.0-fc43.aarch64.rpm |
| Fedora 44 (x86_64) | lemonade-server-11.9.0-fc44.x86_64.rpm |
| Fedora 44 (ARM64) | lemonade-server-11.9.0-fc44.aarch64.rpm |
| macOS | Lemonade-11.9.0-Darwin.pkg |

Other platforms? See our Installation Options for Docker, Snap, Arch, Debian, and more.
Embeddable Lemonade
Portable binaries for bundling into your own installer. Run lemond ./ as a subprocess.

| Platform | Download |
| ------------ | ---------------------------------------------- |
| Ubuntu x64 | lemonade-embeddable-11.9.0-ubuntu-x64.tar.gz |
| Ubuntu arm64 | lemonade-embeddable-11.9.0-ubuntu-arm64.tar.gz |
| Windows x64 | lemonade-embeddable-11.9.0-windows-x64.zip |
| macOS arm64 | lemonade-embeddable-11.9.0-macos-arm64.tar.gz |

What's Changed
Thanks @AaronStGeorge, @Bekhouche, @GabrielReusRodriguez, @Geramy, @abn, @anditherobot, @bitgamma, @fl0rianr, @github-actions, @jeremyfowers, @kenvandine, @superm1, @vaibhav8a, @zaneni6 for your awesome contributions to this release!
Click to expand changelog
- Update FLM version to 1.0.3 by @zaneni6 in https://github.com/lemonade-sdk/lemonade/pull/3381
- Add checkpoints to bench output by @bitgamma in https://github.com/lemonade-sdk/lemonade/pull/3359
- feat(logging): add log rotation and transition default server posture to stdout/stderr by @abn in https://github.com/lemonade-sdk/lemonade/pull/2864
- server: fix MTP draft companion handling by @fl0rianr in https://github.com/lemonade-sdk/lemonade/pull/3368
- docs: fix typo maintainance → maintenance by @vaibhav8a in https://github.com/lemonade-sdk/lemonade/pull/3383
- fix: recover cache and models lost by 11.8.0 FHS split (#3387) by @superm1 in https://github.com/lemonade-sdk/lemonade/pull/3393
- ci(sdcpp): run the CPU validation leg only on schedule/workflow_dispatch by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/3406
- test(gpu-hang): scale request timeouts to the model-operation budget by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/3366
- feat(backends): hrx backend [experimental] by @AaronStGeorge in https://github.com/lemonade-sdk/lemonade/pull/3374
- fix(routing): llm classifier no longer treats has_tools as risk by @Bekhouche in https://github.com/lemonade-sdk/lemonade/pull/3330
- fix(test): pre-download the model in test_gpu_hang_retry before the timed load by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/3408
- fix(server): classify extra models using reserved directories: chat, embeddings, reranking instead of trea

Headline
- Systemd upgrades now recover config and downloaded backends and keep already-downloaded models in place, fixing the 11.8.0 FHS-split regression that misplaced them.
For any users that tried to recover from breakage; please see https://github.com/lemonade-sdk/lemonade/wiki/Migration#v117x--v1180---v1181 for guidance.
Breaking Changes
- The systemd HF_HOME=/var/cache/huggingface redirect was removed, so HuggingFace models now resolve under /var/lib/lemonade/.cache/huggingface (HOME-based) instead of /var/cache/huggingface.
- Ubuntu 25.10 packaging was dropped now that it is EOL and is no longer a supported package target.
Lemonade Server

| Operating System | Downloads |
| ------------------ | ----------------------------------------- |
| Windows | lemonade.msi |
| Ubuntu 24.04+ | Launchpad PPA |
| Debian 13 (x86_64) | lemonade-server_11.8.1-debian13_amd64.deb |
| Debian 13 (ARM64) | lemonade-server_11.8.1-debian13_arm64.deb |
| Fedora 43 (x86_64) | lemonade-server-11.8.1-fc43.x86_64.rpm |
| Fedora 43 (ARM64) | lemonade-server-11.8.1-fc43.aarch64.rpm |
| Fedora 44 (x86_64) | lemonade-server-11.8.1-fc44.x86_64.rpm |
| Fedora 44 (ARM64) | lemonade-server-11.8.1-fc44.aarch64.rpm |
| macOS | Lemonade-11.8.1-Darwin.pkg |

Other platforms? See our Installation Options for Docker, Snap, Arch, Debian, and more.
Embeddable Lemonade
Portable binaries for bundling into your own installer. Run lemond ./ as a subprocess.

| Platform | Download |
| ------------ | ---------------------------------------------- |
| Ubuntu x64 | lemonade-embeddable-11.8.1-ubuntu-x64.tar.gz |
| Ubuntu arm64 | lemonade-embeddable-11.8.1-ubuntu-arm64.tar.gz |
| Windows x64 | lemonade-embeddable-11.8.1-windows-x64.zip |
| macOS arm64 | lemonade-embeddable-11.8.1-macos-arm64.tar.gz |

Full Changelog: https://github.com/lemonade-sdk/lemonade/compare/v11.8.0...v11.8.1

Windows installers are signed. Free code signing provided by SignPath.io, certificate by SignPath Foundation. See our Code Signing Policy.

Headline
- A new POST /v1/models/register endpoint registers or updates user.* model definitions without downloading any files.
- New GET/POST/DELETE /v1/models/{id}/options endpoints save, inspect, and reset per-model recipe options without loading the model, with per-load transient overrides.
- New GET /v1/stats and Prometheus /metrics endpoints report prefix-cache effectiveness and router route-switch counters.
- The built-in catalog adds the Qwen3.8-27B and NVIDIA Nemotron 3.5 Lightning 30B-A3B GGUF models and the Z-Image-Turbo image model.
- Lemonade can now be installed on Windows with winget and on macOS with a Homebrew cask.
Breaking Changes
- POST /pull now returns 400 and registers nothing when a model definition names an unservable or contradictory deployment mode; such labels are no longer silently normalized.
- Stored user_models.json entries that violate the new deployment-mode label rule are now skipped at startup with an error instead of being repaired.
- Removed the deprecated Lite Collection and Ultra Collection model registry entries.
- /v1/audio/transcriptions now returns raw text bodies for text/srt/vtt instead of a JSON {text: ...} wrapper, and returns 400 for unsupported response_format values.
- The TheNoise image backend no longer accepts per-request upscale and lora_dir recipe options; lora_dir is now set server-wide in config.json and upscale is only honored as a passed-through request parameter.
- lemonade launch opencode now emits a limit object (limit.context/limit.output) instead of a top-level contextWindow field.
- The no_broadcast config key is replaced by an inverted broadcast boolean, auto-migrated on load.
- An existing but unreadable extra_models_dir now returns 400 on /internal/set instead of being silently applied.
Lemonade Server

| Operating System | Downloads |
| ------------------ | ----------------------------------------- |
| Windows | lemonade.msi |
| Ubuntu 24.04+ | Launchpad PPA |
| Debian 13 (x86_64) | lemonade-server_11.7.0-debian13_amd64.deb |
| Debian 13 (ARM64) | lemonade-server_11.7.0-debian13_arm64.deb |
| Fedora 43 (x86_64) | lemonade-server-11.7.0-fc43.x86_64.rpm |
| Fedora 43 (ARM64) | lemonade-server-11.7.0-fc43.aarch64.rpm |
| Fedora 44 (x86_64) | lemonade-server-11.7.0-fc44.x86_64.rpm |
| Fedora 44 (ARM64) | lemonade-server-11.7.0-fc44.aarch64.rpm |
| macOS | Lemonade-11.7.0-Darwin.pkg |

Other platforms? See our Installation Options for Docker, Snap, Arch, Debian, and more.
Embeddable Lemonade
Portable binaries for bundling into your own installer. Run lemond ./ as a subprocess.

| Platform | Download |
| ------------ | ---------------------------------------------- |
| Ubuntu x64 | lemonade-embeddable-11.7.0-ubuntu-x64.tar.gz |
| Ubuntu arm64 | lemonade-embeddable-11.7.0-ubuntu-arm64.tar.gz |
| Windows x64 | lemonade-embeddable-11.7.0-windows-x64.zip |
| macOS arm64 | lemonade-embeddable-11.7.0-macos-arm64.tar.gz |

What's Changed
Thanks @SlawomirNowaczyk, @abn, @anditherobot, @bitgamma, @blackdeathdrow, @bong-water-water-bong, @fl0rianr, @ianbmacdonald, @iswaryaalex, @jeremyfowers, @kenvandine, @mzyy94, @osimarr, @popey, @ramkrishna2910, @superm1, @yelliver, @zaneni6 for your awesome contributions to this release!
Click to expand changelog
- Bump TheNoise to version v0.1.2 by @bitgamma in https://github.com/lemonade-sdk/lemonade/pull/3079
- docs(backends): drop the model catalog tables from the backend reference by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/3103
- fix(alias): report the real reason for alias validation failures by @ramkrishna2910 in https://github.com/lemonade-sdk/lemonade/pull/3111
- feat(server): add model registration endpoint by @fl0rianr in https://github.com/lemonade-sdk/lemonade/pull/3118
- Debian packaging improvements by @superm1 in https://github.com/lemonade-sdk/lemonade/pull/3039
- docs: add winget and Homebrew install to homepage quickstart by @yelliver in https://github.com/lemonade-sdk/lemonade/pull/2970
- Trigg

Headline
- The Muse-Glimmer-30B-GGUF model joins the built-in llama.cpp catalog with draft decoding, vision, and tool-calling support.
- The llama.cpp ROCm backend now runs on AMD Instinct MI100, MI200, MI210, and MI250 GPUs on Linux.
- A new experimental TheNoise ROCm image-generation backend brings Anima and Krea-2 to AMD Strix Halo and Strix Point iGPUs.
- The new lemonade alias commands let you assign your own names to models, with alias-aware /v1/models listing and admin endpoints.
- lemonade bench adds opt-in vision benchmark scenarios for multimodal models.
Breaking Changes
- /v1/audio/speech with stream_format=audio and a non-PCM response_format (such as mp3) against Kokoro now returns HTTP 400 (only pcm is supported) instead of silently returning PCM audio.
- /v1/models now lists each GGUF quant in a split extra_models_dir subfolder as its own entry rather than one entry per folder; the old folder name still resolves as a hidden request alias.
- The Windows MSI layout changed: Lemonade_Server_MSI now contains only lemonade-server-minimal.msi, and the desktop lemonade.msi now ships in a new Lemonade_Desktop_MSI artifact, so automation that reads lemonade.msi from the old artifact must be repointed.
Lemonade Server

| Operating System | Downloads |
| ------------------ | ----------------------------------------- |
| Windows | lemonade.msi |
| Ubuntu 24.04+ | Launchpad PPA |
| Debian 13 (x86_64) | lemonade-server_11.6.0-debian13_amd64.deb |
| Debian 13 (ARM64) | lemonade-server_11.6.0-debian13_arm64.deb |
| Fedora 43 (x86_64) | lemonade-server-11.6.0-fc43.x86_64.rpm |
| Fedora 43 (ARM64) | lemonade-server-11.6.0-fc43.aarch64.rpm |
| Fedora 44 (x86_64) | lemonade-server-11.6.0-fc44.x86_64.rpm |
| Fedora 44 (ARM64) | lemonade-server-11.6.0-fc44.aarch64.rpm |
| macOS | Lemonade-11.6.0-Darwin.pkg |

Other platforms? See our Installation Options for Docker, Snap, Arch, Debian, and more.
Embeddable Lemonade
Portable binaries for bundling into your own installer. Run lemond ./ as a subprocess.

| Platform | Download |
| ------------ | ---------------------------------------------- |
| Ubuntu x64 | lemonade-embeddable-11.6.0-ubuntu-x64.tar.gz |
| Ubuntu arm64 | lemonade-embeddable-11.6.0-ubuntu-arm64.tar.gz |
| Windows x64 | lemonade-embeddable-11.6.0-windows-x64.zip |
| macOS arm64 | lemonade-embeddable-11.6.0-macos-arm64.tar.gz |

What's Changed
Thanks @Bekhouche, @SlawomirNowaczyk, @Yigtwxx, @ZMXJJ, @abn, @anditherobot, @bitgamma, @blackdeathdrow, @ckuethe, @duggiefresh, @fl0rianr, @github-actions, @hogeheer499-commits, @jeremyfowers, @kenvandine, @popey, @ramkrishna2910, @sreeram-11, @storm1er, @superm1 for your awesome contributions to this release!
Click to expand changelog
- fix(ci): prevent MSI verify flake from stale ARP entries on shared runners by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/2895
- docs: add testing guide by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/2891
- fix(build): patch libwebsockets GENHDR OUTPUT to silence MSB8065 by @blackdeathdrow in https://github.com/lemonade-sdk/lemonade/pull/2893
- ci: stop re-running the full suite on push to main by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/2899
- fix(cli): respect explicit --port when --host includes a scheme by @kenvandine in https://github.com/lemonade-sdk/lemonade/pull/2914
- README.md - Fix download links by @duggiefresh in https://github.com/lemonade-sdk/lemonade/pull/2827
- ci: run packaging, PPA, distro, macOS and backend-validation jobs in the merge queue by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/2911
- New helper combining add_test() and register_cpp_ci_test() by @SlawomirNowaczyk in https://github.com/lemonade-sdk/lemonade/pull/2877
- [llamacpp] replace deprecated flags, better IO flag handling by @bitgamma in https://github.com/lemonade-sdk/lemonade/pull/2833
- Add /rerank as the reranking endpoint path (alias /reranking) by @sreeram-11 in https://github.com/l

Headline
- The standalone macOS/Linux lemonade-tray client can now connect to servers over HTTPS/TLS.
- Repeatable llama.cpp custom arguments such as --override-kv now preserve every occurrence instead of collapsing to the last value.
- lemonade bench output now includes a hardware section covering CPU, GPU, RAM, OS, and backends.
- The llama.cpp backend now returns a clear error when an explicit device prefix contradicts the resolved backend.
- The Windows MSI installer now restores a saved custom per-user install location during upgrades.
Breaking Changes
Lemonade Server

| Operating System | Downloads |
| ------------------ | ----------------------------------------- |
| Windows | lemonade.msi |
| Ubuntu 24.04+ | Launchpad PPA |
| Debian 13 (x86_64) | lemonade-server_11.5.2-debian13_amd64.deb |
| Debian 13 (ARM64) | lemonade-server_11.5.2-debian13_arm64.deb |
| Fedora 43 (x86_64) | lemonade-server-11.5.2-fc43.x86_64.rpm |
| Fedora 43 (ARM64) | lemonade-server-11.5.2-fc43.aarch64.rpm |
| Fedora 44 (x86_64) | lemonade-server-11.5.2-fc44.x86_64.rpm |
| Fedora 44 (ARM64) | lemonade-server-11.5.2-fc44.aarch64.rpm |
| macOS | Lemonade-11.5.2-Darwin.pkg |

Other platforms? See our Installation Options for Docker, Snap, Arch, Debian, and more.
Embeddable Lemonade
Portable binaries for bundling into your own installer. Run lemond ./ as a subprocess.

| Platform | Download |
| ------------ | ---------------------------------------------- |
| Ubuntu x64 | lemonade-embeddable-11.5.2-ubuntu-x64.tar.gz |
| Ubuntu arm64 | lemonade-embeddable-11.5.2-ubuntu-arm64.tar.gz |
| Windows x64 | lemonade-embeddable-11.5.2-windows-x64.zip |
| macOS arm64 | lemonade-embeddable-11.5.2-macos-arm64.tar.gz |

What's Changed
Thanks @abn, @bitgamma, @ckuethe, @fl0rianr, @github-actions, @jeremyfowers, @kenvandine, @meghsat, @sujikathir for your awesome contributions to this release!
Click to expand changelog
- fix(test): stabilize semantic router similarity thresholds by @fl0rianr in https://github.com/lemonade-sdk/lemonade/pull/2824
- ci: add and scope concurrency policies across workflows by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/2851
- feat(tray): support HTTPS server configurations by @abn in https://github.com/lemonade-sdk/lemonade/pull/2815
- ci: improve persistent Windows runner hygiene and process lock recovery by @abn in https://github.com/lemonade-sdk/lemonade/pull/2837
- fix(build): fix system cpp-httplib detection and HTTPS compilation by @abn in https://github.com/lemonade-sdk/lemonade/pull/2814
- fix(custom_args): preserve repeatable flags by @sujikathir in https://github.com/lemonade-sdk/lemonade/pull/2840
- fix(installer): remember custom install location by @sujikathir in https://github.com/lemonade-sdk/lemonade/pull/2855
- test(ci): wait for server after host rebind by @fl0rianr in https://github.com/lemonade-sdk/lemonade/pull/2870
- ci: make C++ test selection merge-friendly by @fl0rianr in https://github.com/lemonade-sdk/lemonade/pull/2829
- [bench] Refactor & add hardware info by @bitgamma in https://github.com/lemonade-sdk/lemonade/pull/2866
- fix(llamacpp): validate device-backend compatibility during model load by @abn in https://github.com/lemonade-sdk/lemonade/pull/2839
- document multi checkpoint model creation by @ckuethe in https://github.com/lemonade-sdk/lemonade/pull/2844
- FIX(ci-test-router): recover semantic routing after transient embed failure by @fl0rianr in https://github.com/lemonade-sdk/lemonade/pull/2871
- Moving from Quick Rules/ Advanced Rules to Natural Language Router now gives a warning that the progress will be cleared by @meghsat in https://github.com/lemonade-sdk/lemonade/pull/2867
- fix(docs): no support for llama.cpp:cpu, metal is used by @fl0rianr in https://github.com/lemonade-sdk/lemonade/pull/2881
- Update llama.cpp to b10241 by @github-actions[bot] in https://github.com/lemonade-sdk/lemonade/pull/2884
- Bump project version from 11.5.1 to 11.5.2 by @kenvandine in https://github

Headline
- lemonade bench adds an image-generation benchmark mode with capability-aware scenario and model filtering, response capture, and a new --timeout flag.
- The Router Builder gains a Test Prompt tab, backed by a new POST /routing/validate endpoint, that runs a routing policy against a sample prompt and shows the decision, a step-by-step trace, and a decision-tree view.
- Ten MiniCPM text and vision GGUF models from ModelScope join the built-in llama.cpp catalog.
- Tool-calling requests to llama.cpp with large JSON schema bounds are now accepted, working around a grammar limit that previously rejected valid tool calls.
Breaking Changes
- llama.cpp non-streaming responses now return the requested or registered model id in the model field instead of the local .gguf absolute path.
- Registered and imported collections (user.*/extra.*) now list under their canonical prefixed id on /v1/models, Ollama /api/tags, and MCP lemonade_list_models.
- auto backend selection now prefers Vulkan on AMD systems across all engines, matching the llamacpp behavior.
Lemonade Server

| Operating System | Downloads |
| ------------------ | ----------------------------------------- |
| Windows | lemonade.msi |
| Ubuntu 24.04+ | Launchpad PPA |
| Debian 13 (x86_64) | lemonade-server_11.5.1-debian13_amd64.deb |
| Debian 13 (ARM64) | lemonade-server_11.5.1-debian13_arm64.deb |
| Fedora 43 (x86_64) | lemonade-server-11.5.1-fc43.x86_64.rpm |
| Fedora 43 (ARM64) | lemonade-server-11.5.1-fc43.aarch64.rpm |
| Fedora 44 (x86_64) | lemonade-server-11.5.1-fc44.x86_64.rpm |
| Fedora 44 (ARM64) | lemonade-server-11.5.1-fc44.aarch64.rpm |
| macOS | Lemonade-11.5.1-Darwin.pkg |

Other platforms? See our Installation Options for Docker, Snap, Arch, Debian, and more.
Embeddable Lemonade
Portable binaries for bundling into your own installer. Run lemond ./ as a subprocess.

| Platform | Download |
| ------------ | ---------------------------------------------- |
| Ubuntu x64 | lemonade-embeddable-11.5.1-ubuntu-x64.tar.gz |
| Ubuntu arm64 | lemonade-embeddable-11.5.1-ubuntu-arm64.tar.gz |
| Windows x64 | lemonade-embeddable-11.5.1-windows-x64.zip |
| macOS arm64 | lemonade-embeddable-11.5.1-macos-arm64.tar.gz |

What's Changed
Thanks @QuMuzafferEge, @QuTosh, @SlawomirNowaczyk, @ZMXJJ, @abn, @ckuethe, @coolduderules, @fl0rianr, @jeremyfowers, @kenvandine, @ramkrishna2910, @sujikathir, @superm1 for your awesome contributions to this release!
Click to expand changelog
- Update trellis checksums to the pinned v0.4.3 assets by @ramkrishna2910 in https://github.com/lemonade-sdk/lemonade/pull/2778
- web: community links in navbar hamburger + homepage hero by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/2784
- Bench imagegen by @ckuethe in https://github.com/lemonade-sdk/lemonade/pull/2683
- Add libgomp runtime dependency to deb/rpm packages by @kenvandine in https://github.com/lemonade-sdk/lemonade/pull/2794
- Fix: Repeated model update reports after pull by @fl0rianr in https://github.com/lemonade-sdk/lemonade/pull/2715
- [Router] Fix DirectoryWatcher missing policy delete+rewrite on Windows by @SlawomirNowaczyk in https://github.com/lemonade-sdk/lemonade/pull/2798
- Fix llama.cpp model field echoing local .gguf path by @SlawomirNowaczyk in https://github.com/lemonade-sdk/lemonade/pull/2796
- Add JSON schema checking to pre-commit by @superm1 in https://github.com/lemonade-sdk/lemonade/pull/2802
- fix(cli): parse configuration values as JSON arrays and objects by @abn in https://github.com/lemonade-sdk/lemonade/pull/2816
- Fix JSON pre-commit validation with permissive schema by @superm1 with @Copilot in https://github.com/lemonade-sdk/lemonade/pull/2810
- fix(http): add libcurl protocol fallback for macOS libcurl compatibility by @abn in https://github.com/lemonade-sdk/lemonade/pull/2806
- fix(custom_args): treat negative numbers as values, not flags by @coolduderules in https://github.com/lemonade-sdk/lemonade/pull/2773
- Issue 2651: Router Prompt D

Headline
- The Lemonade Router is complete: name a collection.router model and requests are steered by rule, classifier, semantic-similarity, or LLM-as-router policies, with per-candidate decision traces, a GUI for authoring and editing routers, and a POST /v1/classify text-classification endpoint backed by a new onnxruntime backend.
- ModelScope is now searchable alongside Hugging Face in the Model Manager, with source-aware model discovery, download, and registration.
- A server-side job engine lets clients post multi-step recipes and manage them through /jobs endpoints, with pause, interrupt, resume, and delete lifecycle control.
- lemond can act as an MCP client host, connecting to external stdio MCP servers and discovering their tools through admin-gated /internal/mcp/* endpoints.
- The lemonade CLI now supports HTTPS/TLS connections to secure remote endpoints, honoring http(s):// schemes in LEMONADE_HOST.
Breaking Changes
- The server no longer sends Access-Control-Allow-Origin: * by default: non-loopback browser origins are now rejected with 403 unless listed in LEMONADE_ALLOWED_ORIGINS, and the WebSocket same-origin fallback has been removed. Remote or LAN browser clients must set LEMONADE_ALLOWED_ORIGINS.
Lemonade Server

| Operating System | Downloads |
| ------------------ | ----------------------------------------- |
| Windows | lemonade.msi |
| Ubuntu 24.04+ | Launchpad PPA |
| Debian 13 (x86_64) | lemonade-server_11.5.0-debian13_amd64.deb |
| Debian 13 (ARM64) | lemonade-server_11.5.0-debian13_arm64.deb |
| Fedora 43 (x86_64) | lemonade-server-11.5.0-fc43.x86_64.rpm |
| Fedora 43 (ARM64) | lemonade-server-11.5.0-fc43.aarch64.rpm |
| Fedora 44 (x86_64) | lemonade-server-11.5.0-fc44.x86_64.rpm |
| Fedora 44 (ARM64) | lemonade-server-11.5.0-fc44.aarch64.rpm |
| macOS | Lemonade-11.5.0-Darwin.pkg |

Other platforms? See our Installation Options for Docker, Snap, Arch, Debian, and more.
Embeddable Lemonade
Portable binaries for bundling into your own installer. Run lemond ./ as a subprocess.

| Platform | Download |
| ------------ | ---------------------------------------------- |
| Ubuntu x64 | lemonade-embeddable-11.5.0-ubuntu-x64.tar.gz |
| Ubuntu arm64 | lemonade-embeddable-11.5.0-ubuntu-arm64.tar.gz |
| Windows x64 | lemonade-embeddable-11.5.0-windows-x64.zip |
| macOS arm64 | lemonade-embeddable-11.5.0-macos-arm64.tar.gz |

What's Changed
Thanks @OussamaBATOUCHE, @QuManu, @SlawomirNowaczyk, @abn, @ckopanos, @ekkoitac, @fl0rianr, @jon-hotaisle, @kenvandine, @kyuubyN, @meghsat, @midu16, @pwilkin, @radhaksri, @ramkrishna2910, @superm1, @therealjuanmartinez for your awesome contributions to this release!
Click to expand changelog
- Update version to prepare for release by @superm1 in https://github.com/lemonade-sdk/lemonade/pull/2702
- fix(security): pin SHA-256 checksums for contributor-hosted backend assets by @superm1 in https://github.com/lemonade-sdk/lemonade/pull/2710
- fix(ci): always run macOS C++ unit tests regardless of signing by @superm1 in https://github.com/lemonade-sdk/lemonade/pull/2712
- docs (fix): typo in readme by @jon-hotaisle in https://github.com/lemonade-sdk/lemonade/pull/2701
- feat(router): onnxruntime classification backend + /v1/classify (capability for #2592; policy wiring in #2384) by @ramkrishna2910 in https://github.com/lemonade-sdk/lemonade/pull/2626
- Fix: re-pulling a collection.router model fails with "unknown key 'source'" by @SlawomirNowaczyk in https://github.com/lemonade-sdk/lemonade/pull/2703
- feat: add Modelscope search by @fl0rianr in https://github.com/lemonade-sdk/lemonade/pull/2716
- fix(websocket): allow Windows Tauri WebView2 origin for upgrades by @superm1 in https://github.com/lemonade-sdk/lemonade/pull/2721
- fix: quote /MANIFESTINPUT path and detect RX 7600 as gfx110X by @kyuubyN in https://github.com/lemonade-sdk/lemonade/pull/2678
- Add server-side MCP client host foundation by @fl0rianr in https://github.com/lemonade-sdk/lemonade/pull/2605
- fix(docker): Add li

Headline
- A new 3D-generation modality ports the Trellis.2 image-to-3D pipeline, exposing a POST /v1/3d/generations endpoint and a GUI 3D panel with an in-app model viewer.
- Text-to-speech arrives with the OpenMOSS backend, including voice-cloning and voice-design models and a dedicated TTS panel in the GUI.
- Naming a collection.router model on /chat/completions, /completions, or /responses now runs its routing engine and reports the selected route via an x_lemonade_route body field and x-lemonade-route header.
- ModelScope is now supported as a second remote model registry alongside Hugging Face, with source selection across the CLI, API, and desktop app.
- The FastFlowLM NPU backend now auto-installs on Linux instead of requiring a manual system-package install, with a new prefer_system config flag.
Breaking Changes
- The Docker runtime image now runs lemond as unprivileged UID 10001 and stores caches under /opt/lemonade/.cache instead of /root/.cache; chown or recreate existing root-owned named volumes.
- Re-registering a Hugging Face user model under a name already used by a different checkpoint now returns an error instead of silently overwriting the existing user.* model; choose a different name.
Lemonade Server
Operating System Downloads
Windows lemonade.msi
Ubuntu 24.04+ Launchpad PPA
Debian 13 (x86_64) lemonade-server_11.0.0-debian13_amd64.deb
Debian 13 (ARM64) lemonade-server_11.0.0-debian13_arm64.deb
Fedora 43 (x86_64) lemonade-server-11.0.0-fc43.x86_64.rpm
Fedora 43 (ARM64) lemonade-server-11.0.0-fc43.aarch64.rpm
Fedora 44 (x86_64) lemonade-server-11.0.0-fc44.x86_64.rpm
Fedora 44 (ARM64) lemonade-server-11.0.0-fc44.aarch64.rpm
macOS Lemonade-11.0.0-Darwin.pkg
Other platforms? See our Installation Options for Docker, Snap, Arch, Debian, and more.
Embeddable Lemonade
Portable binaries for bundling into your own installer. Run lemond ./ as a subprocess.
Platform Download
Ubuntu x64 lemonade-embeddable-11.0.0-ubuntu-x64.tar.gz
Ubuntu arm64 lemonade-embeddable-11.0.0-ubuntu-arm64.tar.gz
Windows x64 lemonade-embeddable-11.0.0-windows-x64.zip
macOS arm64 lemonade-embeddable-11.0.0-macos-arm64.tar.gz

What's Changed
Thanks @SlawomirNowaczyk, @abn, @ckuethe, @eddierichter-amd, @fl0rianr, @ianbmacdonald, @indistinctTalk, @kenvandine, @kyuubyN, @pwilkin, @ramkrishna2910, @superm1, @wariobot09 for your awesome contributions to this release!
Click to expand changelog
- Update FLM to 0.9.44 and enable portable FLM for Linux by @superm1 in https://github.com/lemonade-sdk/lemonade/pull/2544
- dev: change base image to Ubuntu 24.04 in Dockerfile by @fl0rianr in https://github.com/lemonade-sdk/lemonade/pull/2578
- fix(server): replace dummy bind with connect check for port availability by @abn in https://github.com/lemonade-sdk/lemonade/pull/2576
- Router/2385 collection.router recipe dispatch by @SlawomirNowaczyk in https://github.com/lemonade-sdk/lemonade/pull/2558
- feat(telemetry): add WebSocket server endpoint and security controls by @abn in https://github.com/lemonade-sdk/lemonade/pull/2574
- feat(rocm): enable AMD Instinct MI300X (gfx942) in the llama.cpp ROCm backend by @ianbmacdonald in https://github.com/lemonade-sdk/lemonade/pull/2587
- Asset generation: OpenMOSS by @pwilkin in https://github.com/lemonade-sdk/lemonade/pull/2580
- Start an abstraction layer for benchmark types by @ckuethe in https://github.com/lemonade-sdk/lemonade/pull/2557
- Asset generation: Trellis by @pwilkin in https://github.com/lemonade-sdk/lemonade/pull/2581
- fix(backends): enable sd-cpp CUDA support on Windows by @kenvandine in https://github.com/lemonade-sdk/lemonade/pull/2601
- refactor(audio-gen): apply backend review feedback to ThinkSound and ACE-Step by @pwilkin in https://github.com/lemonade-sdk/lemonade/pull/2594
- fix(rocm): detect version-suffixed HIP runtime on Windows (ROCm 7.x) by @ramkrishna2910 in https://github.com/lemonade-sdk/lemonade/pull/2603
- Router/2386 decision trace response by @eddierichter-amd in https://github.com

Headline
- Audio generation arrives with the ThinkSound sound-effects and ACE-Step music backends, a new POST /v1/audio/generations endpoint, and a GUI audio-generation panel.
- System-installed ROCm is now reused on Windows as well as Linux, and the experimental vLLM ROCm backend reliably targets AMD GPUs on hybrid AMD/NVIDIA hosts.
- A suspend inhibitor keeps the machine awake while inference is running on Linux (systemd-logind), enabled by default and opt-out.
- A new GET /v1/models/{id}/files endpoint reports each model's resolved local files for the GUI Files tab, with absolute paths gated behind an opt-in include_paths=true.
Breaking Changes
- The experimental ROCm vLLM backend now requires a writable $XDG_RUNTIME_DIR; launch fails if one cannot be created, which may affect minimal or container setups.
- After the Debian/Ubuntu package upgrade migration runs, re-enable the per-user service manually with systemctl --user enable lemond.
Lemonade Server
Operating System Downloads
Windows lemonade.msi
Ubuntu 24.04+ Launchpad PPA
Debian 13 (x86_64) lemonade-server_10.10.0-debian13_amd64.deb
Debian 13 (ARM64) lemonade-server_10.10.0-debian13_arm64.deb
Fedora 43 (x86_64) lemonade-server-10.10.0-fc43.x86_64.rpm
Fedora 43 (ARM64) lemonade-server-10.10.0-fc43.aarch64.rpm
Fedora 44 (x86_64) lemonade-server-10.10.0-fc44.x86_64.rpm
Fedora 44 (ARM64) lemonade-server-10.10.0-fc44.aarch64.rpm
macOS Lemonade-10.10.0-Darwin.pkg
Other platforms? See our Installation Options for Docker, Snap, Arch, Debian, and more.
Embeddable Lemonade
Portable binaries for bundling into your own installer. Run lemond ./ as a subprocess.
Platform Download
Ubuntu x64 lemonade-embeddable-10.10.0-ubuntu-x64.tar.gz
Ubuntu arm64 lemonade-embeddable-10.10.0-ubuntu-arm64.tar.gz
Windows x64 lemonade-embeddable-10.10.0-windows-x64.zip
macOS arm64 lemonade-embeddable-10.10.0-macos-arm64.tar.gz

What's Changed
Thanks @SlawomirNowaczyk, @abn, @blackdeathdrow, @coolduderules, @eddierichter-amd, @fl0rianr, @jeremyfowers, @kenvandine, @pwilkin, @sofiageo, @superm1, @volen-silo, @wariobot09 for your awesome contributions to this release!
Click to expand changelog
- feat(rocm): detect externally-installed ROCm on Windows by @volen-silo in https://github.com/lemonade-sdk/lemonade/pull/2508
- Router/2382 RoutingPolicyEngine — assembly + evaluate() (M8) by @SlawomirNowaczyk in https://github.com/lemonade-sdk/lemonade/pull/2499
- docs: add Telemetry guide to mkdocs nav by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/2514
- fix(vLLM): ensures vLLM sees only ROCm devices by @fl0rianr in https://github.com/lemonade-sdk/lemonade/pull/2007
- docs: update arch linux instructions by @sofiageo in https://github.com/lemonade-sdk/lemonade/pull/2517
- fix(installer): add missing architecture_defaults.json to MSI by @blackdeathdrow in https://github.com/lemonade-sdk/lemonade/pull/2525
- fix(app): store Windows favicon as regular file by @fl0rianr in https://github.com/lemonade-sdk/lemonade/pull/2526
- Add support for inhibiting suspend while inference is running by @superm1 in https://github.com/lemonade-sdk/lemonade/pull/2515
- fix: llama.cpp backend updates make sure artifacts are present by @fl0rianr in https://github.com/lemonade-sdk/lemonade/pull/2518
- feat: add model files endpoint by @fl0rianr in https://github.com/lemonade-sdk/lemonade/pull/2437
- docs: cover tests and API docs in adding-a-backend guide by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/2541
- Fix: flaky Whisper realtime transcription event handling by @fl0rianr in https://github.com/lemonade-sdk/lemonade/pull/2535
- Fix disabling systemd user service by @superm1 in https://github.com/lemonade-sdk/lemonade/pull/2516
- fix: support nested schemas and handle client disconnects in responses telemetry by @abn in https://github.com/lemonade-sdk/lemonade/pull/2532
- feat(telemetry): reduce verbosity of telemetry logging by @abn in https://github.com/lemonade-sdk/lemonade/pull/2545
- fix(

Headline
- ARM64 (aarch64) Linux packages now ship for Debian, Fedora, the portable embeddable archive, and Docker.
- Users can set custom system prompts for Lemonade omni models, as demonstrated in the new RPG-HaloTales-V1 narrative experience model.
- A new OTLP telemetry subsystem exports traces using OpenInference and OpenTelemetry GenAI conventions, driven by a lemonade telemetry command.
- The vLLM backend now works with Claude Code through the Anthropic Messages API, and adds Qwen3.6 and GLM-4.7-Flash models.
Breaking Changes
- The Linux system tray is now disabled by default (REQUIRE_LINUX_TRAY=OFF), so Linux server packages ship headless; rebuild with the tray option enabled to restore it.
Lemonade Server
Operating System Downloads
Windows lemonade.msi
Ubuntu 24.04+ Launchpad PPA
Debian 13 (x86_64) lemonade-server_10.9.0-debian13_amd64.deb
Debian 13 (ARM64) lemonade-server_10.9.0-debian13_arm64.deb
Fedora 43 (x86_64) lemonade-server-10.9.0-fc43.x86_64.rpm
Fedora 43 (ARM64) lemonade-server-10.9.0-fc43.aarch64.rpm
Fedora 44 (x86_64) lemonade-server-10.9.0-fc44.x86_64.rpm
Fedora 44 (ARM64) lemonade-server-10.9.0-fc44.aarch64.rpm
macOS Lemonade-10.9.0-Darwin.pkg
Other platforms? See our Installation Options for Docker, Snap, Arch, Debian, and more.
Embeddable Lemonade
Portable binaries for bundling into your own installer. Run lemond ./ as a subprocess.
Platform Download
Ubuntu x64 lemonade-embeddable-10.9.0-ubuntu-x64.tar.gz
Ubuntu arm64 lemonade-embeddable-10.9.0-ubuntu-arm64.tar.gz
Windows x64 lemonade-embeddable-10.9.0-windows-x64.zip
macOS arm64 lemonade-embeddable-10.9.0-macos-arm64.tar.gz

What's Changed
Thanks @GabrielReusRodriguez, @SlawomirNowaczyk, @abn, @bitgamma, @blackdeathdrow, @bong-water-water-bong, @ckuethe, @clemperorpenguin, @eddierichter-amd, @fl0rianr, @gchriswill, @github-actions, @ianbmacdonald, @jeremyfowers, @kenvandine, @lucifer-vali, @ramkrishna2910, @sawansri, @siavashhub, @superm1, @volen-silo for your awesome contributions to this release!
Click to expand changelog
- fix(packaging): don't auto-enable the per-user lemond.service (Closes #2125) by @ianbmacdonald in https://github.com/lemonade-sdk/lemonade/pull/2354
- fix(config): accept flat backend keys like vllm_args in config set (#1824) by @ramkrishna2910 in https://github.com/lemonade-sdk/lemonade/pull/2394
- test/ci: harden CLI endpoint suites against shared-server flakes by @fl0rianr in https://github.com/lemonade-sdk/lemonade/pull/2365
- Lemonade cli list models with size in Gb by @GabrielReusRodriguez in https://github.com/lemonade-sdk/lemonade/pull/2329
- Handle separate sd-cpp CUDA release pins by @fl0rianr in https://github.com/lemonade-sdk/lemonade/pull/2369
- fix: refuse to start a second lemond when the port is already in use (#2255) by @siavashhub in https://github.com/lemonade-sdk/lemonade/pull/2258
- Fix sd-cpp:metal install (resolve asset filename wildcard) by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/2402
- Allow HTTP custom backends with warnings by @sawansri in https://github.com/lemonade-sdk/lemonade/pull/2370
- Add comment and PR description guidelines to AGENTS.md by @superm1 in https://github.com/lemonade-sdk/lemonade/pull/2409
- [Router] Interfaces + fixtures stub (foundation) by @ramkrishna2910 in https://github.com/lemonade-sdk/lemonade/pull/2408
- fix: Prevent the Rosetta prompt on macOS installation by @clemperorpenguin in https://github.com/lemonade-sdk/lemonade/pull/2297
- fix(cli): resolve option inheritance bug in launch help options by @abn in https://github.com/lemonade-sdk/lemonade/pull/2396
- Move omni collection system prompt to per-model JSON by @ramkrishna2910 in https://github.com/lemonade-sdk/lemonade/pull/2395
- fix: ROCm Support for RDNA2 Device Models by @gchriswill in https://github.com/lemonade-sdk/lemonade/pull/2410
- feat: expand system stats with storage infos by @fl0rianr in https://github.com/lemonade-sdk/lemonade/pull/2436
- test: fix whisper realtime macOS VAD flake b

Headline
- Speculative decoding now accepts draft/MTP/EAGLE3 checkpoints, with new Gemma-4 MTP models ready to pull.
- ROCm installation and GPU detection are restored for Radeon RX RDNA2/3/4 dGPUs on Windows and Linux.
- Backend installs are now crash-safe, with resilient downloads that fall back gracefully when a release lookup or cache snapshot is unavailable.
- lemonade bench --response-log captures each model's responses and run metadata to a JSONL file for later quality evaluation.
- The lemonade backends command now lists only supported recipes and backends by default; use lemonade backends --all to see every available option.
Breaking Changes
- The --model-draft, -md, and --spec-draft-model flags are now reserved for internal speculation-decoding support and can no longer be passed manually through llamacpp_args.
Lemonade Server
Operating System Downloads
Windows lemonade.msi
Ubuntu 24.04+ Launchpad PPA
Debian 13 lemonade-server_10.8.1-debian13_amd64.deb
Fedora 43 lemonade-server-10.8.1-fc43.x86_64.rpm
Fedora 44 lemonade-server-10.8.1-fc44.x86_64.rpm
macOS Lemonade-10.8.1-Darwin.pkg
Other platforms? See our Installation Options for Docker, Snap, Arch, Debian, and more.
Embeddable Lemonade
Portable binaries for bundling into your own installer. Run lemond ./ as a subprocess.
Platform Download
Ubuntu x64 lemonade-embeddable-10.8.1-ubuntu-x64.tar.gz
Windows x64 lemonade-embeddable-10.8.1-windows-x64.zip
macOS arm64 lemonade-embeddable-10.8.1-macos-arm64.tar.gz

What's Changed
Thanks @GabrielReusRodriguez, @Kushal1213, @Phqen1x, @abn, @bitgamma, @blackdeathdrow, @ckuethe, @fl0rianr, @github-actions, @ianbmacdonald, @jeremyfowers, @jtlayton, @kenvandine, @lucifer-vali, @matthewjhunter, @ramkrishna2910, @sagebind, @superm1 for your awesome contributions to this release!
Click to expand changelog
- ci: support tagging releases from a release branch by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/2272
- ci: add repo-manager workflow by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/2276
- fix(whisper): drop gfx103X from rocm whisper supported archs by @ramkrishna2910 in https://github.com/lemonade-sdk/lemonade/pull/2274
- test(gguf): add unit tests for MTP / capability label detection (#2176) by @ramkrishna2910 in https://github.com/lemonade-sdk/lemonade/pull/2281
- fix(rocm): accept wildcard GPU arch families in TheRock install gate (#2093 follow-up) by @ramkrishna2910 in https://github.com/lemonade-sdk/lemonade/pull/2280
- fix(test): require Whisper model load before language transcription test by @fl0rianr in https://github.com/lemonade-sdk/lemonade/pull/2278
- docs(cli): correct launch note - LEMONADE_* recipe env vars no longer honored by @ramkrishna2910 in https://github.com/lemonade-sdk/lemonade/pull/2275
- ci(repo-manager): automatic repo-manager fully operational by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/2285
- docs(release): update release process for repo-manager automation by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/2288
- fix(ci): restart MacOS server before whisper metal tests by @fl0rianr in https://github.com/lemonade-sdk/lemonade/pull/2303
- auto update and validate sd-cpp by @fl0rianr in https://github.com/lemonade-sdk/lemonade/pull/2139
- fix(macOS): updates whisper from v1.8.4 to v.1.8.5 for metal by @fl0rianr in https://github.com/lemonade-sdk/lemonade/pull/2309
- Fix Problem with .devcontainer. It did not copy .devcontainer/reinstall-cmake.sh when building container by @GabrielReusRodriguez in https://github.com/lemonade-sdk/lemonade/pull/2273
- Add support for additional draft checkpoint by @bitgamma in https://github.com/lemonade-sdk/lemonade/pull/2317
- docs: fix Debian 13 installation docs for issue #2299 by @superm1 in https://github.com/lemonade-sdk/lemonade/pull/2323
- docs: add documentation style guide for community and AI-assisted contributions by @kenvandine in https://github.com/lemonade-sdk/lemonade/pull/2054
- Fi

Headline
- Lemonade's true omni-modal features have gotten a glow-up!
- LMX-Omni-52B-Halo and LMX-Omni-5.5B-Lite replace the old "collections".
- Add your own custom omni models in the GUI or CLI.
- Downloading custom models have been improved with auto-detected sizes, labels, and improved download robustness.
Lemonade Server
Operating System Downloads
Windows lemonade.msi
Ubuntu 24.04+ Launchpad PPA
Fedora 43 lemonade-server-10.6.0-fc43.x86_64.rpm
Fedora 44 lemonade-server-10.6.0-fc44.x86_64.rpm
macOS Lemonade-10.6.0-Darwin.pkg
Other platforms? See our Installation Options for Docker, Snap, Arch, Debian, and more.
Embeddable Lemonade
Portable binaries for bundling into your own installer. Run lemond ./ as a subprocess.
Platform Download
Ubuntu x64 lemonade-embeddable-10.6.0-ubuntu-x64.tar.gz
Windows x64 lemonade-embeddable-10.6.0-windows-x64.zip
macOS arm64 lemonade-embeddable-10.6.0-macos-arm64.tar.gz

What's Changed
Thanks @Geramy, @abn, @fl0rianr, @github-actions, @jeremyfowers, @neubig, @sofiageo, @superm1 for your awesome contributions to this release!
Click to expand changelog
- fix(sd-cpp): update macOS metal asset filename by @Geramy in https://github.com/lemonade-sdk/lemonade/pull/1949
- Fix : Add 'vision' label to Qwen MTP models by @fl0rianr in https://github.com/lemonade-sdk/lemonade/pull/1952
- d/copyright: Update licensing by @superm1 in https://github.com/lemonade-sdk/lemonade/pull/1955
- Preserve backend error status by @neubig in https://github.com/lemonade-sdk/lemonade/pull/1933
- fix(download manager): restore model download state across renderer r… by @fl0rianr in https://github.com/lemonade-sdk/lemonade/pull/1948
- Add a tool for updating model sizes by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/1962
- packaging: Linux systemd/sysusers compliance for downstream distros by @abn in https://github.com/lemonade-sdk/lemonade/pull/1954
- Register lemonade:// scheme handler in app desktop entry by @abn in https://github.com/lemonade-sdk/lemonade/pull/1957
- fix(app): normalize cleared server base URL to localhost mode by @fl0rianr in https://github.com/lemonade-sdk/lemonade/pull/1961
- fix(app): don't skip port discovery when no explicit base URL is set by @abn in https://github.com/lemonade-sdk/lemonade/pull/1956
- Enable custom user collections via /pull by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/1842
- fix: custom models documentation link by @sofiageo in https://github.com/lemonade-sdk/lemonade/pull/1972
- Improve download UX and correctness by @fl0rianr in https://github.com/lemonade-sdk/lemonade/pull/1764
- Update llama.cpp to b9253 by @github-actions[bot] in https://github.com/lemonade-sdk/lemonade/pull/1975
- Feat : Add UI options for custom Omni Models by @fl0rianr in https://github.com/lemonade-sdk/lemonade/pull/1769
- Rebrand OmniRouter to Lemonade Omni by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/1973
- fix: missing file size in custom models by @sofiageo in https://github.com/lemonade-sdk/lemonade/pull/1888
- feat: added GGUF auto labeling for tools and vision by @fl0rianr in https://github.com/lemonade-sdk/lemonade/pull/1983
- Rev the version number to v10.6.0 by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/1982
- fix: emit collection components under public model names by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/1985
New Contributors
- @neubig made their first contribution in https://github.com/lemonade-sdk/lemonade/pull/1933
Full Changelog: https://github.com/lemonade-sdk/lemonade/compare/v10.5.1...v10.6.0

Windows installers are signed. Free code signing provided by SignPath.io, certificate by SignPath Foundation. See our Code Signing Policy.

Lemonade Server
Operating System Downloads
Windows lemonade.msi
Ubuntu 24.04+ Launchpad PPA
Fedora 43 lemonade-server-10.5.1-fc43.x86_64.rpm
Fedora 44 lemonade-server-10.5.1-fc44.x86_64.rpm
macOS Lemonade-10.5.1-Darwin.pkg
Other platforms? See our Installation Options for Docker, Snap, Arch, Debian, and more.
Embeddable Lemonade
Portable binaries for bundling into your own installer. Run lemond ./ as a subprocess.
Platform Download
Ubuntu x64 lemonade-embeddable-10.5.1-ubuntu-x64.tar.gz
Windows x64 lemonade-embeddable-10.5.1-windows-x64.zip
macOS arm64 lemonade-embeddable-10.5.1-macos-arm64.tar.gz

What's Changed
Thanks @bitgamma, @dmitrii-galantsev, @fl0rianr, @github-actions, @jeremyfowers, @kpoineal, @lucifer-vali, @seii, @sofiageo, @superm1 for your awesome contributions to this release!
Click to expand changelog
- feat(app): add full custom sd.cpp model options by @fl0rianr in https://github.com/lemonade-sdk/lemonade/pull/1909
- fix: update bug report template with current CLI commands by @kpoineal in https://github.com/lemonade-sdk/lemonade/pull/1914
- keep quotes during arg merging by @bitgamma in https://github.com/lemonade-sdk/lemonade/pull/1920
- Separate releases for Fedora 43 and 44 by @lucifer-vali in https://github.com/lemonade-sdk/lemonade/pull/1917
- Update Fedora maintainer in contribution guide by @lucifer-vali in https://github.com/lemonade-sdk/lemonade/pull/1928
- fix: add missing rdna1, rdna2, and gfx1150 targets by @sofiageo in https://github.com/lemonade-sdk/lemonade/pull/1923
- Introduce per-backend args by @bitgamma in https://github.com/lemonade-sdk/lemonade/pull/1915
- Update llama.cpp to b9196 by @github-actions[bot] in https://github.com/lemonade-sdk/lemonade/pull/1931
- Polish auto-generated release notes by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/1935
- Expose llama.cpp "/tokenize" endpoint by @seii in https://github.com/lemonade-sdk/lemonade/pull/1926
- test(app): add regression guards for model options and collections by @fl0rianr in https://github.com/lemonade-sdk/lemonade/pull/1786
- Switch to upstream builds for stable diffusion by @superm1 in https://github.com/lemonade-sdk/lemonade/pull/1922
- fix: Update and remove non-existing .md links by @dmitrii-galantsev in https://github.com/lemonade-sdk/lemonade/pull/1938
- fix: resolve rocm channel when checking expected sd-cpp/llamacpp version by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/1934
- feat: built-in MTP support by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/1944
- remove deprecated lemonade-server CLI by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/1918
- fix: vLLM memory behavior on APUs by @fl0rianr in https://github.com/lemonade-sdk/lemonade/pull/1919
- Update llama.cpp to b9213 by @github-actions[bot] in https://github.com/lemonade-sdk/lemonade/pull/1946
New Contributors
- @lucifer-vali made their first contribution in https://github.com/lemonade-sdk/lemonade/pull/1917
- @seii made their first contribution in https://github.com/lemonade-sdk/lemonade/pull/1926
- @dmitrii-galantsev made their first contribution in https://github.com/lemonade-sdk/lemonade/pull/1938
Full Changelog: https://github.com/lemonade-sdk/lemonade/compare/v10.5.0...v10.5.1

Windows installers are signed. Free code signing provided by SignPath.io, certificate by SignPath Foundation. See our Code Signing Policy.

Headline
- Upgraded to ROCm 7.13 for llama.cpp and stable-diffusion.cpp!
- macOS has graduated from beta with all major features fully supported.
- Overhauled management of custom/imported models and recipes.
Breaking changes
- room-preview has been deprecated because TheRock is now stable! You will be auto-migrated from rocm-preview to rocm-stable
- audio has been renamed to transcription in the /v1//health endpoint
- If you have *.args set in the global config and *_args set for a specific model, these options will be merged at load time. Previously, per-model args would completely replace global args. You can opt-out on a per-model basis by setting merge_args=False in a load request and/or saved model options.
Lemonade Server
Operating System Downloads
Windows lemonade.msi
Ubuntu 24.04+ Launchpad PPA
Fedora 44 lemonade-server-10.5.0.x86_64.rpm
macOS Lemonade-10.5.0-Darwin.pkg
lemonade-server-minimal.msi and lemonade-server CLI are deprecated and will be removed in the next release.
Other platforms? See our Installation Guide for Docker, Snap, Arch, and more.
Embeddable Lemonade
Portable binaries for bundling into your own installer. Run lemond ./ as a subprocess.
Platform Download
Ubuntu x64 lemonade-embeddable-10.5.0-ubuntu-x64.tar.gz
Windows x64 lemonade-embeddable-10.5.0-windows-x64.zip
macOS arm64 lemonade-embeddable-10.5.0-macos-arm64.tar.gz

What's Changed
Thanks @Geramy, @SorenDreano, @Tetramatrix, @bitgamma, @fl0rianr, @genrtul, @github-actions, @gunnarx, @jeremyfowers, @meghsat, @sawansri, @storm1er, @superm1 for your awesome contributions to this release!
Click to expand changelog
- Fix website links by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/1855
- Add the Development docs nav entry for philosophy.md by @Copilot in https://github.com/lemonade-sdk/lemonade/pull/1851
- Device option sep by @genrtul in https://github.com/lemonade-sdk/lemonade/pull/1397
- fix model resolution for sharded models on windows by @bitgamma in https://github.com/lemonade-sdk/lemonade/pull/1868
- Update llama.cpp to b9095 by @github-actions[bot] in https://github.com/lemonade-sdk/lemonade/pull/1861
- feat(setup.sh): add rpm-ostree support for Fedora Atomic/Desktop environments by @storm1er in https://github.com/lemonade-sdk/lemonade/pull/1797
- fix link to the contribution guide in the README.md by @SorenDreano in https://github.com/lemonade-sdk/lemonade/pull/1877
- Dockerfile: libxrt2 & libxrt-npu2 dependencies for fastflowlm by @gunnarx in https://github.com/lemonade-sdk/lemonade/pull/1801
- Let vLLM choose model dtype automatically instead of forcing float16 by @SorenDreano in https://github.com/lemonade-sdk/lemonade/pull/1876
- Fixing opencode unreachable on Windows bug by @meghsat in https://github.com/lemonade-sdk/lemonade/pull/1886
- Set FLM_CONFIG_PATH for CI by @superm1 in https://github.com/lemonade-sdk/lemonade/pull/1889
- Fix: Properly terminate child processes on systemd service stop by @superm1 in https://github.com/lemonade-sdk/lemonade/pull/1866
- Remove stale cli options for lemonade launch by @sawansri in https://github.com/lemonade-sdk/lemonade/pull/1880
- Restore sd-cpp:rocm channel fallback when rocm_channel=nightly by @Copilot in https://github.com/lemonade-sdk/lemonade/pull/1882
- Build web app by default on Windows by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/1891
- Add a getting started guide by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/1895
- Merge arguments instead of replacing them. by @bitgamma in https://github.com/lemonade-sdk/lemonade/pull/1873
- Rev FLM to 0.9.42 and fix transcription labeling system by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/1810
- feat(macos): add macOS arm64 (Metal) support for whisper, kokoros, sd-cpp by @Geramy in https://github.com/lemonade-sdk/lemonade/pull/1777
- Update ROCm to 7.13 by @superm1 in https://github.com/lemonade-sdk/lemonade/pull/1900
- Canonical model IDs with friendly-name precedenc

Lemonade Server
Operating System Downloads
Windows lemonade.msi (Server + App)
Ubuntu 24.04+ Launchpad PPA (Server + App)
Fedora 43 lemonade-server-10.4.0.x86_64.rpm (Server + App)
macOS (beta) Lemonade-10.4.0-Darwin.pkg (Server + App)
lemonade-server-minimal.msi is deprecated and will be removed in a future release.
Other platforms? See our Installation Options for Docker, Snap, Arch, Fedora, Debian, and more.
Embeddable Lemonade
Portable binaries for bundling into your own installer. Run lemond ./ as a subprocess.
Platform Download
Ubuntu x64 lemonade-embeddable-10.4.0-ubuntu-x64.tar.gz
Windows x64 lemonade-embeddable-10.4.0-windows-x64.zip

What's Changed
Thanks @amd-pworfolk, @anditherobot, @jeremyfowers, @osimarr, @ramkrishna2910, @superm1, @theonlychant for your awesome contributions to this release!
Click to expand changelog
- chore(linux): remove the AppImage build target by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/1756
- docs: add llamacpp.md to nav and document rocm_channel in configuration reference by @Copilot in https://github.com/lemonade-sdk/lemonade/pull/1755
- Npm security by @superm1 in https://github.com/lemonade-sdk/lemonade/pull/1665
- Restructure the documentation by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/1776
- Support /slots endpoint for llamacpp backend by @amd-pworfolk in https://github.com/lemonade-sdk/lemonade/pull/1722
- feat: integrate Qwen2.5-Omni models with native audio input support by @ramkrishna2910 in https://github.com/lemonade-sdk/lemonade/pull/1581
- New contribution guide and philosophy document by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/1793
- Fix macOS dev build: drop Linux-only libdrm and ad-hoc codesign default by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/1796
- Add vLLM as a wrapped server backend (ROCm) by @ramkrishna2910 in https://github.com/lemonade-sdk/lemonade/pull/1537
- Update the docs onto the new design system by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/1798
- Replace model label dots with icons by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/1800
- Improve model manager naming and grouping by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/1803
- Fix backend → backends typo in llamacpp docs by @Copilot in https://github.com/lemonade-sdk/lemonade/pull/1808
- docs for vLLM backend by @ramkrishna2910 in https://github.com/lemonade-sdk/lemonade/pull/1802
- vllm: wire --ctx-size and --vllm-args end-to-end by @ramkrishna2910 in https://github.com/lemonade-sdk/lemonade/pull/1827
- Add static max context metadata to /models by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/1823
- Warm model cache asynchronously by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/1829
- fix(logs): build WebSocket URL from full server base URL by @theonlychant in https://github.com/lemonade-sdk/lemonade/pull/1837
- docs: clarify that llamacpp.rocm_bin version tags are channel-specific by @Copilot in https://github.com/lemonade-sdk/lemonade/pull/1815
- vllm: fix BackendManager download progress for split-archive installs by @ramkrishna2910 in https://github.com/lemonade-sdk/lemonade/pull/1832
- Add vLLM ROCm news article by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/1844
- Expose wrapped backend process IDs in /health by @anditherobot in https://github.com/lemonade-sdk/lemonade/pull/1838
- Fix ollama tool calling by @osimarr in https://github.com/lemonade-sdk/lemonade/pull/1780
New Contributors
- @theonlychant made their first contribution in https://github.com/lemonade-sdk/lemonade/pull/1837
Full Changelog: https://github.com/lemonade-sdk/lemonade/compare/v10.3.0...v10.4.0

Windows installers are signed. Free code signing provided by SignPath.io, certificate by SignPath Foundation. See our Code Signing Policy.

Headline
- OmniRouter unifies the best backend engines to give you true omni-modal LLM chat. Make and edit images, generate speech, and more with natural language. Open AI-compliant.
- The desktop app is now based on Tauri instead of Electron, so it is 10x smaller.
- Easily update to specific llama.cpp versions, or enable auto-update.
- The GUI has a light mode theme!
- ROCm 7.2 stable release, 7.12 preview release, and TheRock nightly builds are all supported. Preview is now the default.
Breaking Changes
- lemonade_server.service is now lemond.service.
- The amd_igpu and amd_dgpu fields in system-info have been consolidated to amd_gpu. nvidia_dgpu renamed to nvidia_gpu for consistency.
- ROCm channel default changed from nightly to preview.
Lemonade Server
Operating System Downloads
Windows lemonade.msi (Server + App)
Ubuntu 24.04+ Launchpad PPA (Server) · lemonade-app-10.3.0-x86_64.AppImage (Companion Desktop App)
Fedora 43 lemonade-server-10.3.0.x86_64.rpm (Server) · lemonade-app-10.3.0-x86_64.AppImage (Companion Desktop App)
macOS (beta) Lemonade-10.3.0-Darwin.pkg (Server + App)
lemonade-server-minimal.msi is deprecated and will be removed in a future release.
Other platforms? See our Installation Options for Docker, Snap, Arch, Fedora, Debian, and more.
Embeddable Lemonade
Portable binaries for bundling into your own installer. Run lemond ./ as a subprocess.
Platform Download
Ubuntu x64 lemonade-embeddable-10.3.0-ubuntu-x64.tar.gz
Windows x64 lemonade-embeddable-10.3.0-windows-x64.zip

What's Changed
Thanks @Geramy, @agreenbhm, @anditherobot, @bitgamma, @danielholanda, @fl0rianr, @github-actions, @ianbmacdonald, @jeremyfowers, @martinlroth, @michaelburch, @muriloime, @osimarr, @ramkrishna2910, @sawansri, @siavashhub, @sofiageo, @superm1 for your awesome contributions to this release!
Click to expand changelog
- Fix broken debug build on Linux by @agreenbhm in https://github.com/lemonade-sdk/lemonade/pull/1593
- [codex] add appear-builtin model aliases by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/1590
- fix: skip ROCm download for gfx1033/1035/1036 by @martinlroth in https://github.com/lemonade-sdk/lemonade/pull/1596
- Remove completely stale code.md doc by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/1609
- Update llama.cpp to b8766 by @github-actions[bot] in https://github.com/lemonade-sdk/lemonade/pull/1619
- feat: add light theme by @sofiageo in https://github.com/lemonade-sdk/lemonade/pull/1618
- Fix Claude Code Guide by @sawansri in https://github.com/lemonade-sdk/lemonade/pull/1620
- extend load timeout in CLI by @bitgamma in https://github.com/lemonade-sdk/lemonade/pull/1621
- Disable realtime VAD for manual commit mode by @michaelburch in https://github.com/lemonade-sdk/lemonade/pull/1607
- Remove stale link by @ramkrishna2910 in https://github.com/lemonade-sdk/lemonade/pull/1628
- Migrate desktop app from Electron to Tauri by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/1611
- Add qwen3.6-35b-a3b by @sawansri in https://github.com/lemonade-sdk/lemonade/pull/1650
- Rename lemonade-server.service to lemond.service by @superm1 in https://github.com/lemonade-sdk/lemonade/pull/1651
- Fix FLM version check on Linux to accept installed version >= minimum required by @Copilot in https://github.com/lemonade-sdk/lemonade/pull/1652
- fix: make /load declarative/idempotent to prevent TOCTOU race by @ianbmacdonald in https://github.com/lemonade-sdk/lemonade/pull/1604
- ROCm-Backend: added and tested docker image functionality with ROCm by @siavashhub in https://github.com/lemonade-sdk/lemonade/pull/1587
- Drop AMD iGPU vs dGPU designations by @superm1 in https://github.com/lemonade-sdk/lemonade/pull/1648
- fix: use wss:// for WebSocket connections when served over HTTPS (#1583) by @Geramy in https://github.com/lemonade-sdk/lemonade/pull/1666
- Add a new configuration option to prevent installation of executables by @superm1 in https://github.com/lemonade-sdk/lemonade/pu

Lemonade Server
Operating System Downloads
Windows lemonade.msi (Server + App) · lemonade-server-minimal.msi (Server Only)
Ubuntu 24.04+ Launchpad PPA (Server) · lemonade-app-10.2.0-x86_64.AppImage (Companion Desktop App)
Fedora 43 lemonade-server-10.2.0.x86_64.rpm (Server) · lemonade-app-10.2.0-x86_64.AppImage (Companion Desktop App)
macOS (beta) Lemonade-10.2.0-Darwin.pkg (Server + App)
Other platforms? See our Installation Options for Docker, Snap, Arch, Fedora, Debian, and more.
Embeddable Lemonade
Portable binaries for bundling into your own installer. Run lemond ./ as a subprocess.
Platform Download
Ubuntu x64 lemonade-embeddable-10.2.0-ubuntu-x64.tar.gz
Windows x64 lemonade-embeddable-10.2.0-windows-x64.zip

What's Changed
Thanks @bconsolvo, @bitgamma, @ianbmacdonald, @jeremyfowers, @kenvandine, @ramkrishna2910, @sawansri for your awesome contributions to this release!
Click to expand changelog
- follow up for LEMONADE_ADMIN_API_KEY by @bitgamma in https://github.com/lemonade-sdk/lemonade/pull/1564
- Add embeddable lemonade release artifacts by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/1508
- Overhaul the front README by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/1550
- Add a user guide for Claude Code integration by @sawansri in https://github.com/lemonade-sdk/lemonade/pull/1334
- correct mmproj pull examples by @bitgamma in https://github.com/lemonade-sdk/lemonade/pull/1571
- Fix llamacpp and whispercpp reporting incorrect device types in stats by @kenvandine in https://github.com/lemonade-sdk/lemonade/pull/1558
- Prepare to release v10.1.1 by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/1574
- Update RAM requirement for gpt-oss-20b-FLM filter by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/1570
- feat(tray): add Max Loaded Models submenu by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/1572
- remove duplicate http headers by @bitgamma in https://github.com/lemonade-sdk/lemonade/pull/1573
- fix(server): return empty string when exact GGUF variant not found by @ianbmacdonald in https://github.com/lemonade-sdk/lemonade/pull/1502
- Upgrade flm version to 0.9.38 by @ramkrishna2910 in https://github.com/lemonade-sdk/lemonade/pull/1475
- fix: auto-detect GPU backend for ESRGAN upscaling by @bconsolvo in https://github.com/lemonade-sdk/lemonade/pull/1576
- Add OpenCode integration with launch by @sawansri in https://github.com/lemonade-sdk/lemonade/pull/1516
- Smart CLI pull command by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/1580
- feat: overhaul multi-checkpoint storage strategy and add Qwen Image with additional sd-cpp params by @ianbmacdonald in https://github.com/lemonade-sdk/lemonade/pull/1412
Full Changelog: https://github.com/lemonade-sdk/lemonade/compare/v10.1.0...v10.2.0

Windows installers are signed. Free code signing provided by SignPath.io, certificate by SignPath Foundation. See our Code Signing Policy.

Headline
- The new lemonade CLI is a much nicer way to interact with the service.
- lemonade config replaces most uses of the conf file on Linux.
- Coding is supercharged with improved lemonade launch codex|claude.
- Default port number changed to 13305.
- Support for Gemma 4 models on GPU.
Breaking Changes
Please see the migration guide for this release to best understand the port, CLI, and backend changes in this "spring cleaning" release.
The lemonade-server CLI is deprecated and will be removed in the next release after May 6, 2026.
Quick Install
Operating System Downloads
Windows lemonade.msi (Server + App) · lemonade-server-minimal.msi (Server Only)
Ubuntu 24.04+ Launchpad PPA (Server) · lemonade-app-10.1.0-x86_64.AppImage (Companion Desktop App)
Fedora 43 lemonade-server-10.1.0.x86_64.rpm (Server) · lemonade-app-10.1.0-x86_64.AppImage (Companion Desktop App)
macOS (beta) Lemonade-10.1.0-Darwin.pkg (Server + App)
Other platforms? See our Installation Options for Docker, Snap, Arch, Fedora, Debian, and more.

What's Changed
Thanks @Phqen1x, @bconsolvo, @bitgamma, @ckuethe, @github-actions, @jeremyfowers, @kenvandine, @kincl, @kpoineal, @kronenpj, @kurtseebauer, @ramkrishna2910, @ryomo, @sawansri, @superm1 for your awesome contributions to this release!
Click to expand changelog
- Update FLM NPU directions by @superm1 in https://github.com/lemonade-sdk/lemonade/pull/1430
- Spring cleaning 1 by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/1400
- Update agents.md by @ramkrishna2910 in https://github.com/lemonade-sdk/lemonade/pull/1436
- Point people to Launchpad PPA in GH release by @superm1 in https://github.com/lemonade-sdk/lemonade/pull/1437
- Sort tray includes alphabetically by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/1452
- Rename lemonade-router executable to lemond by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/1451
- Marketplace Submission: Lemon Zest Guide by @Phqen1x in https://github.com/lemonade-sdk/lemonade/pull/1408
- Adding support in backend and frontend UI for super-resolution (Real-ESRGAN models) by @bconsolvo in https://github.com/lemonade-sdk/lemonade/pull/1176
- Fix: GET /api/v1/system-info returns 500 on non-English Windows locales by @ryomo in https://github.com/lemonade-sdk/lemonade/pull/1438
- Add GitHub issue templates by @kpoineal in https://github.com/lemonade-sdk/lemonade/pull/1309
- Correct lemonade recipes command by @superm1 in https://github.com/lemonade-sdk/lemonade/pull/1461
- fix: add PCI revision 0x20 (Krackan Point) to XDNA2 NPU detection by @kurtseebauer in https://github.com/lemonade-sdk/lemonade/pull/1465
- Fix direct-run log streaming for lemonade-router by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/1453
- feat: add openSUSE Tumbleweed support by @kurtseebauer in https://github.com/lemonade-sdk/lemonade/pull/1466
- fix: replace deprecated lemonade-server in distro build tests by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/1480
- Clarify NPU support for Linux in FAQ by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/1477
- Clarify message for unsupported AMD GPUs by @superm1 in https://github.com/lemonade-sdk/lemonade/pull/1467
- Make AppImage build opt-in via BUILD_APPIMAGE option by @kenvandine in https://github.com/lemonade-sdk/lemonade/pull/1387
- Speed up the Windows CI build by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/1455
- Account for HF folder names by @ramkrishna2910 in https://github.com/lemonade-sdk/lemonade/pull/1479
- Set the StartupWMClass by default by @superm1 in https://github.com/lemonade-sdk/lemonade/pull/1492
- test: add environment variable integration tests by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/1491
- Image Edit Feature (Fixes #1352) by @Phqen1x in https://github.com/lemonade-sdk/lemonade/pull/1439
- Make lemonade available in container by @kronenpj in https://github.com/lemonade-sdk/lemonade/pull/148

Headline
- Debian packages are now available via PPA instead of .deb
- Streamlined experience for searching and adding GGUFs from Hugging Face in the app
- qwen3.5-4b now available on NPU via FastFlowLM v0.9.36
- llama.cpp uplifted to b8460 with significant Qwen3.5 optimizations
- Smoother FLM on Linux install process
Ubuntu Installation Change
New installation process for Ubuntu, replacing the .deb installer:
sudo add-apt-repository ppa:lemonade-team/stable
sudo apt install lemonade-server
Quick Install
────────────────┬─────────────────────────────────────────────────────────────────────────────────────────────────────────────
Operating System│Downloads
────────────────┼─────────────────────────────────────────────────────────────────────────────────────────────────────────────
Windows │lemonade.msi (Server + App) · lemonade-server-minimal.msi (Server Only)
────────────────┼─────────────────────────────────────────────────────────────────────────────────────────────────────────────
Fedora 43 │lemonade-server-10.0.1.x86_64.rpm (Server) · Lemonade-10.0.1-x86_64.AppImage (Companion Desktop App)
────────────────┼─────────────────────────────────────────────────────────────────────────────────────────────────────────────
macOS (beta) │Lemonade-10.0.1-Darwin.pkg (Server + App)
────────────────┴─────────────────────────────────────────────────────────────────────────────────────────────────────────────
> Other platforms? See our Installation Options for Docker, Snap, Arch, Fedora, Debian, and more.
What's Changed
Thanks @Rebreda, @abn, @bitgamma, @choppu, @de-wim, @elijahsgh, @github-actions, @ianbmacdonald, @jaeiclee, @jeremyfowers, @kenvandine, @kpoineal, @meghsat, @ramkrishna2910, @sawansri, @sofiageo, @superm1, @szkocot, @timothycarambat for your awesome contributions to this release!
Click to expand changelog
- Recommend ffmpeg in Debian package for whisper audio conversion by @ianbmacdonald in #1325
- Disable microphone when running on insecure Window by @superm1 in #1305
- feat: add whispercpp custom args config by @ianbmacdonald in #1310
- Update flm_npu_linux.html by @de-wim in #1344
- AnythingLLM Docs update by @timothycarambat in #1345
- fix: ignore stdout from flm version --json to ensure we only get valid json by @kenvandine in #1351
- Add libatomic1 as Docker runtime dependency for ROCm backend by @jaeiclee in #1349
- Update flm_npu_linux.html by @de-wim in #1350
- Attempt to preserve capabilities by @superm1 in #1339
- Add conf.d support for Lemonade config overrides by @ianbmacdonald in #1324
- Update NPU Exclusivity constraints in server_spec.md by @elijahsgh in #1356
- fix: fastflowlm guide for Arch Linux by @sofiageo in #1359
- Update NPU exclusivity section in server_spec.md by @jeremyfowers in #1360
- add standalone CLI by @bitgamma in #1336
- Update supported configs table on the readme by @jeremyfowers in #1362
- Fix streaming: call sink.done() on all error paths to properly terminate chunked responses by @abn in #1340
- fix(tray): use get_environment_variable_utf8 for HF cache paths by @kpoineal in #1311
- Add a model form revamp by @meghsat in #1355
- Add query parameter handling to enable firefox chatbot integration by @Rebreda in #1327
- Update CLAUDE.md for v10.0.0 by @ramkrishna2910 in #1383
- Add a new target to build the web app using system nodejs by @superm1 in #1369
- Add Linux system tray support via AppIndicator3 by @abn in #1300
- Create containers for building native packages by @superm1 in #1367
- Fix merge queue CI failures by running apt-get update before deb install by @jeremyfowers in #1393
- Disable hardware caching for Linux by @superm1 in #1354
- Changes to prepare for native packaging by @superm1 in #1366
- Restricting the HF search to only GGUF text gen models by @meghsat in #1392
- Move to libwebsockets by @ramkrishna2910 in #1384
- AGENTS.md for non-claude agents by @sawansri in #1396
- Add libwebsockets-dev to Ubuntu PPA emulation containers by @superm1 in #1402
- Install FLM in th

Headline
- Linux NPU support is available for LLMs and Whisper via FastFlowLM
- Native integration with Claude Code: lemonade-server launch claude
- A Fedora .rpm installer is now published in the release
Quick Install
Operating System Downloads
Windows lemonade.msi (Server + App) · lemonade-server-minimal.msi (Server Only)
Ubuntu lemonade-server_10.0.0_amd64.deb (Server) · Lemonade-10.0.0-x86_64.AppImage (Standalone App)
Fedora lemonade-server-10.0.0.x86_64.rpm (Server) · Lemonade-10.0.0-x86_64.AppImage (Standalone App)
macOS (beta) Lemonade-10.0.0-Darwin.pkg (Server + App)
Other platforms? See our Installation Options for Docker, Snap, Arch, Fedora, Debian, and more.

What's Changed
Thanks @Geramy, @Jays-1111, @OmerFarukOruc, @Rebreda, @abn, @bitgamma, @choppu, @james-martinez, @jeremyfowers, @kenvandine, @ramkrishna2910, @sawansri, @sofiageo, @superm1, @szkocot, @vgodsoe for your awesome contributions to this release!
Click to expand changelog
- Add tool-calling label to more models by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/1224
- Update FLM NPU min kernel version to 6.17 by @superm1 in https://github.com/lemonade-sdk/lemonade/pull/1228
- Fix code block formatting in NPU Linux documentation by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/1238
- Add an NPU linux beta flag by @superm1 in https://github.com/lemonade-sdk/lemonade/pull/1244
- Enable NPU for Linux by @superm1 in https://github.com/lemonade-sdk/lemonade/pull/1243
- Fix false 'not running' status in sandboxed Linux environments by @abn in https://github.com/lemonade-sdk/lemonade/pull/1249
- Fix journald detection using sd_pid_get_unit() instead of JOURNAL_STREAM by @abn in https://github.com/lemonade-sdk/lemonade/pull/1247
- Enable gpt-oss-20b-flm on linux by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/1251
- Fix incorrect skipping of PID file and instance locks by @abn in https://github.com/lemonade-sdk/lemonade/pull/1246
- doc: update Arch Linux NPU instructions by @sofiageo in https://github.com/lemonade-sdk/lemonade/pull/1253
- Use $XDG_RUNTIME_DIR/lemonade for runtime artifacts on Linux by @abn in https://github.com/lemonade-sdk/lemonade/pull/1254
- Fix server failing to start when device is offline by @abn in https://github.com/lemonade-sdk/lemonade/pull/1257
- Fix duplicate server exit messaging by @szkocot in https://github.com/lemonade-sdk/lemonade/pull/1230
- Make Vulkan take priority over CPU for whisper.cpp by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/1252
- Launch flm the same on Windows and Linux by @superm1 in https://github.com/lemonade-sdk/lemonade/pull/1266
- Reduce FLM beta only to flm validate by @superm1 in https://github.com/lemonade-sdk/lemonade/pull/1259
- fix path to to the turbo version of z-image by @bitgamma in https://github.com/lemonade-sdk/lemonade/pull/1264
- Overhaul the blog template by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/1268
- Clarify system requirement on FLM Linux setup by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/1271
- Add support for passing custom arguments to FLM backend through --flm-args by @Jays-1111 in https://github.com/lemonade-sdk/lemonade/pull/1235
- Update vulnerable npm packages by @Rebreda in https://github.com/lemonade-sdk/lemonade/pull/1274
- Enumerate all RFC1918 interfaces for network beacon broadcasting by @Geramy in https://github.com/lemonade-sdk/lemonade/pull/1225
- Make streaming Whisper work in the app when Lemonade server is on another machine by @bitgamma in https://github.com/lemonade-sdk/lemonade/pull/1263
- Add support for NPU Utilization in status bar by @superm1 in https://github.com/lemonade-sdk/lemonade/pull/1261
- Add FLM multi-modal support (ASR + embeddings) with test refactor by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/1270
- Fix Windows Unicode cache paths for downloads by @szkocot in https://github.com/lemonade-sdk/lemonade/pull/1232
- In

Headline
- Support for the Qwen3.5 family of models on ROCm and Vulkan!
- Redesigned the app for easier navigation with a full Backend Manager available in the app, CLI, and endpoints.
- Image generation support is greatly expanded with an image editing endpoint and pull-ing custom models.
- Transcription (whisper.cpp) is available on Linux with Vulkan and CPU support.
Quick Install
Operating System Downloads
Windows lemonade.msi (Server + App) · lemonade-server-minimal.msi (Server Only)
Ubuntu lemonade-server_9.4.1_amd64.deb (Server) · Lemonade-9.4.1-x86_64.AppImage (Standalone App)
macOS (beta) Lemonade-9.4.1-Darwin.pkg (Server + App)
Other platforms? See our Installation Options for Docker, Arch, Fedora, Debian, and more.

What's Changed
Thanks @Geramy, @bitgamma, @iswaryaalex, @jeremyfowers, @ramkrishna2910, @sofiageo, @superm1 for your awesome contributions to this release!
Click to expand changelog
- Update readme by @ramkrishna2910 in https://github.com/lemonade-sdk/lemonade/pull/1174
- Migrate conffile to /etc not /opt/etc by @superm1 in https://github.com/lemonade-sdk/lemonade/pull/1197
- Enable clangd by default by @superm1 in https://github.com/lemonade-sdk/lemonade/pull/1201
- Refactor, fix, and polish the inference UI in the app by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/1195
- Remove unused includes by @bitgamma in https://github.com/lemonade-sdk/lemonade/pull/1202
- Simplify release notes download table layout by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/1205
- Fixes StatusBar loop - split useEffects and updates dependencies by @sofiageo in https://github.com/lemonade-sdk/lemonade/pull/1206
- Fix a few Load Options bugs in the app by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/1204
- Whisper.cpp Linux support for CPU and Vulkan backends by @iswaryaalex in https://github.com/lemonade-sdk/lemonade/pull/1118
- Add more ground work for flm/linux by @superm1 in https://github.com/lemonade-sdk/lemonade/pull/1177
- disable mmap by default if we have iGPU by @bitgamma in https://github.com/lemonade-sdk/lemonade/pull/1211
- Restore gfx1151 kernel check for ROCm work by @superm1 in https://github.com/lemonade-sdk/lemonade/pull/1210
- Update NPU Linux instructions by @superm1 in https://github.com/lemonade-sdk/lemonade/pull/1208
- Add loopback unicast to NetworkBeacon for reliable same-machine disco… by @Geramy in https://github.com/lemonade-sdk/lemonade/pull/1215
- Add background beacon listener for automatic server port discovery by @Geramy in https://github.com/lemonade-sdk/lemonade/pull/1216
- Update NPU references to clarify they're FLM by @superm1 in https://github.com/lemonade-sdk/lemonade/pull/1219
- Update qwen 3.5 model tags by @ramkrishna2910 in https://github.com/lemonade-sdk/lemonade/pull/1220
- Rev llamacpp version for better Qwen3.5 support by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/1214
Full Changelog: https://github.com/lemonade-sdk/lemonade/compare/v9.4.0...v9.4.1

Windows installers are signed. Free code signing provided by SignPath.io, certificate by SignPath Foundation. See our Code Signing Policy.

Headline
- Support for the Qwen3.5 family and LFM2-24B-A2B on llama.cpp + ROCm
- The app's layout has been redesigned, with a new Backend Manager
- Image editing endpoint added
Breaking Change
.deb installation has changed, please see the migration guide.
Other platforms? See our Installation Options for Docker, Arch, Fedora, Debian, and more.

What's Changed
Thanks @Geramy, @ZaneNi, @bitgamma, @danielholanda, @ianbmacdonald, @jeremyfowers, @kenvandine, @sawansri, @superm1 for your awesome contributions to this release!
Click to expand changelog
- Use mDNS beacon to discover server instead of lemonade-server status by @superm1 in https://github.com/lemonade-sdk/lemonade/pull/1158
- Add support for a new appimage target by @superm1 in https://github.com/lemonade-sdk/lemonade/pull/1079
- Add a systemd migration script by @superm1 in https://github.com/lemonade-sdk/lemonade/pull/1078
- Update build/resources/backend_versions.json during cmake by @superm1 in https://github.com/lemonade-sdk/lemonade/pull/1114
- Reduce down to a single Linux package by @superm1 in https://github.com/lemonade-sdk/lemonade/pull/1022
- Update FAQ with Linux STX Halo GPU RAM allocation guidance by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/1170
- Add macOS (beta) mentions across docs and site by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/1165
- Add pulling of image models by @bitgamma in https://github.com/lemonade-sdk/lemonade/pull/1132
- Stop logging to tmpfs when running under systemd by @superm1 in https://github.com/lemonade-sdk/lemonade/pull/1179
- Add unsigned local build+test path for macOS CI when signing secrets … by @Geramy in https://github.com/lemonade-sdk/lemonade/pull/1186
- Add a message for older kernel on gfx1151 by @superm1 in https://github.com/lemonade-sdk/lemonade/pull/1106
- FLM Version Update to v0.9.34 by @ZaneNi in https://github.com/lemonade-sdk/lemonade/pull/1173
- Add Home AI Setup to Open WebUI Docs by @sawansri in https://github.com/lemonade-sdk/lemonade/pull/1180
- Disable system-info CI test on macOS by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/1190
- docs: add --jinja to reserved llamacpp_args list by @ianbmacdonald in https://github.com/lemonade-sdk/lemonade/pull/1037
- Fix macOS race condition in execute_stop_command process checks by @Geramy in https://github.com/lemonade-sdk/lemonade/pull/1183
- feat: Add image edit and variation endpoints by @kenvandine in https://github.com/lemonade-sdk/lemonade/pull/1121
- Prevent pasting images on non-vision models by @danielholanda in https://github.com/lemonade-sdk/lemonade/pull/1164
- App redesign and centralized Backend Manager by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/1171
- Don't hardcode lemonade-router to debug by @superm1 in https://github.com/lemonade-sdk/lemonade/pull/1185
- New llamacpp, Qwen 3.5, and LFM2 24B by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/1192
- Add connection status to the status bar by @superm1 in https://github.com/lemonade-sdk/lemonade/pull/1167
New Contributors
- @sawansri made their first contribution in https://github.com/lemonade-sdk/lemonade/pull/1180
Full Changelog: https://github.com/lemonade-sdk/lemonade/compare/v9.3.4...v9.4.0

Windows installers are signed. Free code signing provided by SignPath.io, certificate by SignPath Foundation. See our Code Signing Policy.

Headline
- macOS (beta) installers are now a part of the release!
- Text-to-speech is now available in the Lemonade App with the Kokoro recipe.
- XDNA 2 NPU detection has been overhauled for better compatibility.
Quick Install
Operating System Server + App Server Only
Windows lemonade.msi lemonade-server-minimal.msi
Ubuntu lemonade_9.3.4_amd64.deb lemonade-server-minimal_9.3.4_amd64.deb
macOS (beta) Lemonade-9.3.4-Darwin.pkg --
Other platforms? See our Installation Options for Docker, Arch, Fedora, Debian, and more.

What's Changed
Thanks @Geramy, @bitgamma, @jeremyfowers, @superm1 for your awesome contributions to this release!
Click to expand changelog
- Add macOS CI/CD: build, sign, notarize, test, and release .pkg by @Geramy in https://github.com/lemonade-sdk/lemonade/pull/1081
- Add support for NPU detection on Linux by @superm1 in https://github.com/lemonade-sdk/lemonade/pull/1091
- Define TTS model type separately from audio by @bitgamma in https://github.com/lemonade-sdk/lemonade/pull/1131
- Fix: webapp not reachable at its normal address by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/1150
- Kokoro UI integration by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/1147
- Optimize CI by 8 minutes by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/1155
- Detect XDNA2 NPU via PCI device ID instead of CPU name regex by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/1154
Full Changelog: https://github.com/lemonade-sdk/lemonade/compare/v9.3.3...v9.3.4

Windows installers are signed. Free code signing provided by SignPath.io, certificate by SignPath Foundation. See our Code Signing Policy.

Headline
- FastFlowLM upgraded to 0.9.33 with Qwen2.5-VL-3B-Instruct support.
- Streaming transcription enabled in the app and with /realtime endpoint.
- Added Ollama API support for better native integration with local AI apps.
- Z-Image and Flux image gen models supported (CPU only for now).
Quick Install
Operating System Server + App Server Only
Windows lemonade.msi lemonade-server-minimal.msi
Ubuntu lemonade_9.3.3_amd64.deb lemonade-server-minimal_9.3.3_amd64.deb
Other platforms? See our Installation Options for Docker, Arch, Fedora, Debian, and more.

What's Changed
Thanks @ZaneNi, @bitgamma, @choppu, @jeremyfowers, @makn87amd, @ramkrishna2910, @vugenti for your awesome contributions to this release!
Click to expand changelog
- Update README for multi-modal by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/1110
- FLM Version Update to v0.9.33 by @ZaneNi in https://github.com/lemonade-sdk/lemonade/pull/1109
- Add support for multiple checkpoints in models by @bitgamma in https://github.com/lemonade-sdk/lemonade/pull/1063
- Remove api_image_generation_rocm.py by @makn87amd in https://github.com/lemonade-sdk/lemonade/pull/1096
- add missing backend external binary vars by @bitgamma in https://github.com/lemonade-sdk/lemonade/pull/1115
- Update maintainer list in README.md by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/1138
- Respect LEMONADE_DISABLE_MODEL_FILTERING in model pull/download by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/1139
- Issue #1082 - Support 256K Context Size in tray app context menu by @vugenti in https://github.com/lemonade-sdk/lemonade/pull/1120
- Document LEMONADE_MAX_LOADED_MODELS environment variable by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/1141
- Fix Electron app port discovery always being skipped by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/1142
- Add Ollama API compatibility layer by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/1130
- Whisper streaming audio by @ramkrishna2910 in https://github.com/lemonade-sdk/lemonade/pull/997
- Make SD model load timeout the same as other engines by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/1144
- Prep 9.3.3 release by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/1146
New Contributors
- @vugenti made their first contribution in https://github.com/lemonade-sdk/lemonade/pull/1120
Full Changelog: https://github.com/lemonade-sdk/lemonade/compare/v9.3.2...v9.3.3

Windows installers are signed. Free code signing provided by SignPath.io, certificate by SignPath Foundation. See our Code Signing Policy.

Headline
- Image generation with ROCm is now support on Windows and Linux
- --max-loaded-models N has been simplified and now enables N models of each type
- Lots of under-the-hood improvements to polish the changes from 9.3.0
Quick Install
Operating System Server + App Server Only
Windows lemonade.msi lemonade-server-minimal.msi
Ubuntu lemonade_9.3.2_amd64.deb lemonade-server-minimal_9.3.2_amd64.deb
Other platforms? See our Installation Options for Docker, Arch, Fedora, Debian, and more.

What's Changed
Thanks @bitgamma, @choppu, @jeremyfowers, @makn87amd, @ramkrishna2910, @sofiageo, @superm1, @vgodsoe for your awesome contributions to this release!
Click to expand changelog
- Update systemd dependency for Arch Linux builds by @sofiageo in https://github.com/lemonade-sdk/lemonade/pull/1067
- Adds LEMONADE_API_KEY in secrets file by @sofiageo in https://github.com/lemonade-sdk/lemonade/pull/1057
- fix a typo that breaks external binary lookup by @bitgamma in https://github.com/lemonade-sdk/lemonade/pull/1062
- add ENV for LEMONADE_MAX_LOADED_MODELS by @bitgamma in https://github.com/lemonade-sdk/lemonade/pull/1064
- Fix /v1/* endpoints returning HTML instead of JSON for OpenAI compati… by @superm1 in https://github.com/lemonade-sdk/lemonade/pull/1069
- Remove some unneeded messages from window.api emulation by @superm1 in https://github.com/lemonade-sdk/lemonade/pull/1059
- handle model start event in chat by @choppu in https://github.com/lemonade-sdk/lemonade/pull/1065
- Add ROCm GPU backend support for stable-diffusion.cpp by @makn87amd in https://github.com/lemonade-sdk/lemonade/pull/980
- Rework --max-loaded-models by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/1086
- Add pagination to fix ci monitoring finding the artifacts by @vgodsoe in https://github.com/lemonade-sdk/lemonade/pull/1080
- Update lemonade server LimitMEMLOCK value by @superm1 in https://github.com/lemonade-sdk/lemonade/pull/1088
- Redo the entire website by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/1099
- Update the website more by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/1100
- 1103-CI-Fix by @ramkrishna2910 in https://github.com/lemonade-sdk/lemonade/pull/1105
- Clean the lemonade/bin folder on upgrade from 9.3.1 or below by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/1102
- Improve SD support in the CLI and app by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/1089
- Correct the path used for the SD ROCm artifacts for Linux by @superm1 in https://github.com/lemonade-sdk/lemonade/pull/1085
New Contributors
- @makn87amd made their first contribution in https://github.com/lemonade-sdk/lemonade/pull/980
Full Changelog: https://github.com/lemonade-sdk/lemonade/compare/v9.3.1...v9.3.2

Windows installers are signed. Free code signing provided by SignPath.io, certificate by SignPath Foundation. See our Code Signing Policy.

Headline
- Windows MSI installers are now code signed. Thanks, SignPath Foundation!
- Ryzen AI SW 1.7 is integrated and fully replaces the prior Ryzen AI SW support.
- Lemonade Marketplace is now included in the app.
Breaking Change
- Please read this before using Ryzen AI SW 1.7 via Lemonade, particularly if you are migrating from a previous version.
Quick Install
Operating System Server + App Server Only
Windows lemonade.msi lemonade-server-minimal.msi
Ubuntu lemonade_9.3.1_amd64.deb lemonade-server-minimal_9.3.1_amd64.deb
Other platforms? See our Installation Options for Docker, Arch, Fedora, Debian, and more.

What's Changed
Thanks @bitgamma, @danielholanda, @eddierichter-amd, @jeremyfowers, @kenvandine, @superm1 for your awesome contributions to this release!
Click to expand changelog
- Redesign lemonade featured apps marketplace by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/1045
- Preserve chat message width when editing by @danielholanda in https://github.com/lemonade-sdk/lemonade/pull/1040
- Refactor backends by @bitgamma in https://github.com/lemonade-sdk/lemonade/pull/1036
- Add badege from the snap store to show the latest version published in stable by @kenvandine in https://github.com/lemonade-sdk/lemonade/pull/1033
- Patching max_tokens in llamacpp to openai server max_completion_tokens by @eddierichter-amd in https://github.com/lemonade-sdk/lemonade/pull/1044
- Fix web-app about dialog by @superm1 in https://github.com/lemonade-sdk/lemonade/pull/1050
- Upgrade to Ryzen AI SW 1.7 by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/1056
- Enable MSI signing via SignPath Foundation by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/909
- Add a redirection to web app for mobile users by @superm1 in https://github.com/lemonade-sdk/lemonade/pull/1060
- Polish the app center panel by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/1061
Full Changelog: https://github.com/lemonade-sdk/lemonade/compare/v9.3.0...v9.3.1

Windows installers are signed. Free code signing provided by SignPath.io, certificate by SignPath Foundation. See our Code Signing Policy.

Headline
- NPU enabled for Whisper models on the /audio/transcription endpoint (Windows only).
- Kokoro TTS is available on the /audio/speech endpoint.
- Desktop and web apps have backend selection, providing a ROCm option in the GUI for the first time.
- Significant performance uplift on AMD devices using llamacpp Vulkan and FastFlowLM.
- A lot more! This was a huge release.
Breaking Changes
- The .deb now installs as a systemd service. See https://github.com/lemonade-sdk/lemonade/issues/1006 for details and migration info.
- The inference_engines key has been removed from the /system-info endpoint response, replaced by a new recipes list.
Quick Install
Operating System Server + App Server Only
Windows lemonade.msi lemonade-server-minimal.msi
Ubuntu lemonade_9.3.0_amd64.deb lemonade-server-minimal_9.3.0_amd64.deb
Other platforms? See our Installation Options for Docker, Arch, Fedora, Debian, and more.

What's Changed
Thanks @Geramy, @LaurentChardin, @bitgamma, @choppu, @dennemark, @iswaryaalex, @jeremyfowers, @kenvandine, @ramkrishna2910, @sofiageo, @superm1 for your awesome contributions to this release!
Click to expand changelog
- feat: Add status bar with server and system stats by @kenvandine in https://github.com/lemonade-sdk/lemonade/pull/927
- Remove all trailing whitespace and apply black to all python files by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/966
- enable /responses endpoint for llamacpp recipes by @bitgamma in https://github.com/lemonade-sdk/lemonade/pull/945
- Add load options menu by @sofiageo in https://github.com/lemonade-sdk/lemonade/pull/880
- Fix link to llm-debate.html in instructions by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/984
- Rename SD label by @ramkrishna2910 in https://github.com/lemonade-sdk/lemonade/pull/986
- fix: Don't attempt to preserve ownership when extracting backends by @kenvandine in https://github.com/lemonade-sdk/lemonade/pull/989
- Add a helper script that uses Ninja build for building by @superm1 in https://github.com/lemonade-sdk/lemonade/pull/991
- Add a new /api/v1/system-stats endpoint and use it by @superm1 in https://github.com/lemonade-sdk/lemonade/pull/970
- Update the Installing Models section of the integration guide by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/993
- Update CMake minimum version to use ... syntax by @superm1 in https://github.com/lemonade-sdk/lemonade/pull/999
- Prevent a rare Windows model load bug by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/979
- llamacpp: also reserve flag aliases/negatives by @bitgamma in https://github.com/lemonade-sdk/lemonade/pull/974
- Create and distribute systemd service by @superm1 in https://github.com/lemonade-sdk/lemonade/pull/1002
- Add more system information into about modal by @superm1 in https://github.com/lemonade-sdk/lemonade/pull/1005
- Windows - fix some issues with buildscripts by @LaurentChardin in https://github.com/lemonade-sdk/lemonade/pull/1001
- Add image generation docs openwebui by @dennemark in https://github.com/lemonade-sdk/lemonade/pull/994
- Add connection settings in Electron app by @choppu in https://github.com/lemonade-sdk/lemonade/pull/1004
- Re-do the system info code and endpoint by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/992
- do not close the same stream multiple times by @bitgamma in https://github.com/lemonade-sdk/lemonade/pull/1014
- Geramy/announce to lan by @Geramy in https://github.com/lemonade-sdk/lemonade/pull/988
- feat: Add copy button to code blocks by @kenvandine in https://github.com/lemonade-sdk/lemonade/pull/969
- Fix systemd service installation for non-/usr prefixes by @superm1 in https://github.com/lemonade-sdk/lemonade/pull/1007
- Add Kokoro TTS backend by @bitgamma in https://github.com/lemonade-sdk/lemonade/pull/1000
- Show --help output if no arguments to lemonade-server by @superm1 in https://github.com/lemonade-sdk/lemonade/pull/1017
- Desktop app: llamacpp

Headline
- Introduces the images/generations endpoint with stablediffusion.cpp as the first image generation engine.
- Compatibility improvements: Debian now supported; Debian, Arch, and Fedora builds tested in CI.
- ryzenai-server and lemonade-eval have moved to their own repos under lemonade-sdk org.
- AMD GPU device detection on Linux overhauled.
Quick Install
Operating System Server + App Server Only
Windows lemonade.msi lemonade-server-minimal.msi
Ubuntu lemonade_9.2.0_amd64.deb lemonade-server-minimal_9.2.0_amd64.deb
Other platforms? See our Installation Options for Docker, Arch, Fedora, Debian, and more.

What's Changed
Thanks @Geramy, @NineBallo, @bitgamma, @jeremyfowers, @ramkrishna2910, @siavashhub, @superm1 for your awesome contributions to this release!
Click to expand changelog
- Improve env var and customization documentation by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/884
- Migrate ryzenai-server into its own repo by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/903
- Reload loaded models if calling /load explicitly by @bitgamma in https://github.com/lemonade-sdk/lemonade/pull/905
- refactor(backends): extract archive extraction utilities to shared ba… by @Geramy in https://github.com/lemonade-sdk/lemonade/pull/910
- unify CLI parsing between router and server by @bitgamma in https://github.com/lemonade-sdk/lemonade/pull/874
- Use standard getaddrinfo instead of `httplibdetailgetaddrinfo_w… by @superm1 in https://github.com/lemonade-sdk/lemonade/pull/911
- Docker always tag latest by @siavashhub in https://github.com/lemonade-sdk/lemonade/pull/906
- Add pre-commit support by @superm1 in https://github.com/lemonade-sdk/lemonade/pull/683
- Detect AMD GPUs without lspci by @superm1 in https://github.com/lemonade-sdk/lemonade/pull/419
- Add mobile app cross links by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/921
- Stable diffusion cpp integration by @ramkrishna2910 in https://github.com/lemonade-sdk/lemonade/pull/828
- Refactor the lemonade dev cli into lemonade-eval by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/904
- Testing overhaul by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/929
- Enable more CLI and endpoint testing by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/935
- Add electron app linking instructions by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/914
- Add AMD GTT detection support by @NineBallo in https://github.com/lemonade-sdk/lemonade/pull/901
- Stable diffusion cpp app integration by @ramkrishna2910 in https://github.com/lemonade-sdk/lemonade/pull/930
- Use system libraries if available by @superm1 in https://github.com/lemonade-sdk/lemonade/pull/913
- Fix CLI connection logic by @bitgamma in https://github.com/lemonade-sdk/lemonade/pull/933
- Add snap installation option for Ubuntu by @Copilot in https://github.com/lemonade-sdk/lemonade/pull/934
- Move lemonade-eval to its own repo by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/938
- Fix some compilation warnings I've seen by @superm1 in https://github.com/lemonade-sdk/lemonade/pull/939
- Rev the version number for release by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/940
- Add arch linux to CI by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/946
- Revert "Fix some compilation warnings I've seen" by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/951
- Upgrade arch job to also build fedora and debian by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/950
- reapply API-key related fixes by @bitgamma in https://github.com/lemonade-sdk/lemonade/pull/949
- Generate fancy release notes by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/952
- Auto-generate fancy release notes by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/959
New Contributors
- @superm1 made their first contribution in https://github.co

Headline
- Support GLM-4.7-Flash-GGUF on ROCm and Vulkan
- Support installing for All Users on Windows and Linux
- Allow an absolute path to go into the CLI pull command
- Rev to FastFlowLM 0.9.27 and support new LFM2.5 models
What's Changed
- Update Docker image version in README by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/825
- Added Linux install instructions by @sofiageo in https://github.com/lemonade-sdk/lemonade/pull/829
- Raise exceptions on lm-eval failures by @ramkrishna2910 in https://github.com/lemonade-sdk/lemonade/pull/827
- Download llamacpp/whisper binaries to cache on linux by @bitgamma in https://github.com/lemonade-sdk/lemonade/pull/833
- Add strix halo dedicated gpu memory to the FAQ by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/842
- Add workflows to test and monitor our runners by @vgodsoe in https://github.com/lemonade-sdk/lemonade/pull/843
- Docs: Add missing POST /api/v1/delete to additional endpoints list by @Copilot in https://github.com/lemonade-sdk/lemonade/pull/853
- Replace hybrid model with device-agnostic model in spec by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/845
- Update maintainers list in README.md by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/856
- add libasound2t64 dependency to full deb package by @ianbmacdonald in https://github.com/lemonade-sdk/lemonade/pull/848
- Pack all recipe options into a class by @bitgamma in https://github.com/lemonade-sdk/lemonade/pull/837
- Fix ROCm Detection for Strix Halo on Ubuntu 24.04 OEM Kernel by @goodtiding5 in https://github.com/lemonade-sdk/lemonade/pull/826
- Allow an absolute path to go into the CLI pull command by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/852
- Add: unversioned /live liveness endpoint & align Docker health check by @siavashhub in https://github.com/lemonade-sdk/lemonade/pull/835
- Doc Docker documentation by @siavashhub in https://github.com/lemonade-sdk/lemonade/pull/854
- Add Perplexica to the apps list by @ItzCrazyKns in https://github.com/lemonade-sdk/lemonade/pull/858
- Add server-based load and bench tools to dev cli by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/850
- remove checkpoint_loaded and context_size from /health and add recipe and options in all_models_loaded by @bitgamma in https://github.com/lemonade-sdk/lemonade/pull/857
- Run npm audit fix on the app by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/861
- fix recipe options creation for newly added models by @bitgamma in https://github.com/lemonade-sdk/lemonade/pull/862
- Add stx-test back in for new ci runners by @vgodsoe in https://github.com/lemonade-sdk/lemonade/pull/864
- Rev FLM to 0.9.26 by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/865
- Add Ubuntu version recommendation by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/870
- Add link anchors to install_options.html by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/871
- FLM version updates: New model (breaking change) by @ZaneNi in https://github.com/lemonade-sdk/lemonade/pull/878
- Update the Roadmap section of the README by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/895
- Install for all users on Windows; Rev FLM and llama by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/863
New Contributors
- @sofiageo made their first contribution in https://github.com/lemonade-sdk/lemonade/pull/829
- @goodtiding5 made their first contribution in https://github.com/lemonade-sdk/lemonade/pull/826
- @ItzCrazyKns made their first contribution in https://github.com/lemonade-sdk/lemonade/pull/858
Full Changelog: https://github.com/lemonade-sdk/lemonade/compare/v9.1.3...v9.1.4

Headline
- Lemonade App can be used remotely over the network
- Lemonade CLI and /load endpoint can save custom load options for each model
- LFM2.5 model enabled via FastFlowLM v0.9.25
- Initial support for publishing Docker images
What's Changed
- Overhaul the FAQ by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/811
- Add FLM server models by @ZaneNi in https://github.com/lemonade-sdk/lemonade/pull/806
- implement save_options by @bitgamma in https://github.com/lemonade-sdk/lemonade/pull/804
- Rev flm to v0.9.25 by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/818
- Added ability to build rpm packages for Fedora users by @SidShetye in https://github.com/lemonade-sdk/lemonade/pull/812
- Docker Images & GitHub Actions Workflow for Lemonade Releases by @siavashhub in https://github.com/lemonade-sdk/lemonade/pull/803
- Extend Lemonade CLI to support FastFlow models by @amd-pworfolk in https://github.com/lemonade-sdk/lemonade/pull/616
- remove all state from the llamacpp backend by @bitgamma in https://github.com/lemonade-sdk/lemonade/pull/801
- Enable the app to access a remote lemonade-server by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/814
- fix server detection on Linux by @bitgamma in https://github.com/lemonade-sdk/lemonade/pull/815
New Contributors
- @SidShetye made their first contribution in https://github.com/lemonade-sdk/lemonade/pull/812
Full Changelog: https://github.com/lemonade-sdk/lemonade/compare/v9.1.2...v9.1.3

Headline
- Create, save, and share custom model recipes: https://lemonade-server.ai/docs/server/server_spec/#per-model-options
- Full NPU and ROCm support for the ROG Ally X handheld and RAI Z2 Extreme chip
- Full compatibility with LM Studio's GGUF library using --extra-models-dir
- FastFlowLM 0.9.24 support with improved ease-of-use
What's Changed
- Update GitHub/Website for early 2026 by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/716
- Add nemotron 3 nano model by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/751
- Touch up the documentation by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/753
- Adding instructions on how to build and run Lemonade CPP in a docker container by @siavashhub in https://github.com/lemonade-sdk/lemonade/pull/723
- Fix index.html responsive design by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/755
- Center website subtitle by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/756
- Add Ryzen AI Z2 support and LEMONADE_DISABLE_MODEL_FILTERING env var by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/764
- Upgrade FLM to 0.9.24 by @ramkrishna2910 in https://github.com/lemonade-sdk/lemonade/pull/781
- Steamline the new user experience by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/788
- add support for recipe options by @bitgamma in https://github.com/lemonade-sdk/lemonade/pull/775
- Lemonade API key by @bitgamma in https://github.com/lemonade-sdk/lemonade/pull/774
- Extra models dir models have an extra. prefix by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/793
- Improve the UX when FLM updates invalidate models by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/796
- Fix logprob calculation for hybrid by @ramkrishna2910 in https://github.com/lemonade-sdk/lemonade/pull/777
- Make some llamacpp-server args optional defaults by @bitgamma in https://github.com/lemonade-sdk/lemonade/pull/763
Full Changelog: https://github.com/lemonade-sdk/lemonade/compare/v9.1.1...v9.1.2

Headline
- Add support for transcriptions, reranking, and embeddings in the App
- ROCm support for Strix Point (gfx1150)
- Solve a bunch of community-reported feedback
What's Changed
- Revamp README file by @danielholanda in https://github.com/lemonade-sdk/lemonade/pull/668
- Update installer guide by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/708
- Add core Lemonade features to homepage by @vgodsoe in https://github.com/lemonade-sdk/lemonade/pull/702
- Fix site icons not loading by @vgodsoe in https://github.com/lemonade-sdk/lemonade/pull/712
- Update wall of logos on the website by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/714
- Resolve cmake warnings by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/713
- Update Runner Labels by @danielholanda in https://github.com/lemonade-sdk/lemonade/pull/728
- Fix bug that resulted in user-model registrations, Add a Model bugs by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/729
- Improve FLM installation and fix bugs by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/730
- Overhaul the "not-found" error message by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/735
- Fix app download progress bugs with 0B files by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/743
- Suppress health request prints by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/742
- Add experiences for Embeddings, Reranking, and Transcription by @danielholanda in https://github.com/lemonade-sdk/lemonade/pull/731
- Add --extra-models-dir option by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/741
- Add ROCm support to STX Point (gfx1150) by @danielholanda in https://github.com/lemonade-sdk/lemonade/pull/740
Full Changelog: https://github.com/lemonade-sdk/lemonade/compare/v9.1.0...v9.1.1

Headline
- The Lemonade App is here! Download models, try them out, and discover the Lemonade ecosystem.
- Initial support for ASR with whisper.cpp.
- Fixes: localhost vs. 127.0.0.1 aliasing; FLM-related issues.
Breaking Changes
- The browser-based web UI has been completely removed.
- A new installer, lemonade.msi / .deb has been introduced, which includes the new Lemonade App.
- The previous lemonade-server-minimal installers still exist but are server-only (no app).
What's Changed
- Fix debate bug with Jan model name by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/666
- Documentation for Debate Arena, and associated models by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/667
- Whisper.cpp Audio Transcription Support by @ramkrishna2910 in https://github.com/lemonade-sdk/lemonade/pull/645
- Lemonade Manager: Phase 2 by @danielholanda in https://github.com/lemonade-sdk/lemonade/pull/647
- Add multi-model support to UI by @danielholanda in https://github.com/lemonade-sdk/lemonade/pull/675
- Dev Containers - Docker build system implemented and includes cmake p… by @Geramy in https://github.com/lemonade-sdk/lemonade/pull/658
- Improve FLM logic and add LFM2-1.2B-FLM by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/693
- Add clear guidance on UI/UX contributions by @danielholanda in https://github.com/lemonade-sdk/lemonade/pull/679
- Refactor LLM chat code maximize code reuse and show FLM thinking box by @danielholanda in https://github.com/lemonade-sdk/lemonade/pull/686
- Fetch the models list when the user clicks on the dropdown by @danielholanda in https://github.com/lemonade-sdk/lemonade/pull/694
- Allow overriding whisper-server bin path via the environment by @bitgamma in https://github.com/lemonade-sdk/lemonade/pull/673
- Add download manager to UI by @danielholanda in https://github.com/lemonade-sdk/lemonade/pull/676
- Setting server host now resolves to ipv4 instead of being ambiguous by @Geramy in https://github.com/lemonade-sdk/lemonade/pull/661
- Expand Model Manager models by default if only one recipe is available by @danielholanda in https://github.com/lemonade-sdk/lemonade/pull/705
- Fix lemonade-server run on Linux by @danielholanda in https://github.com/lemonade-sdk/lemonade/pull/701
- Route index.html to api/v1 by @danielholanda in https://github.com/lemonade-sdk/lemonade/pull/699
- Filter available models by size and system ram by @jeremyfowers in https://github.com/lemonade-sdk/lemonade/pull/696
- Add "Collapse Thinking by Default" Setting by @danielholanda in https://github.com/lemonade-sdk/lemonade/pull/678
New Contributors
- @Geramy made their first contribution in https://github.com/lemonade-sdk/lemonade/pull/658
Full Changelog: https://github.com/lemonade-sdk/lemonade/compare/v9.0.8...v9.1.0

What's Changed
- Set the FLM server host name by @jeremyfowers in #654
- Allow overriding llama-server bin path through the environment by @bitgamma in #651
- Adding CPU Backend Support to LlamaCPP Server for Lemonade CPP by @siavashhub in #627
- Debate Arena v2, and multi-model improvements by @jeremyfowers in #659
New Contributors
- @bitgamma made their first contribution in #651
Full Changelog: https://github.com/lemonade-sdk/lemonade/compare/v9.0.7...v9.0.8