| commit |
1.00 |
cmake: make external compat patches idempotent (#17948) |
|
Commit message contains explicit AI assi |
2026-08-26 |
| commit |
1.00 |
server/images: prevent skipVerify map collision with duplica |
|
Commit message contains explicit AI assi |
2026-08-12 |
| pr |
1.00 |
feat: Add GraniteForCausalLM support in experimental models |
|
Explicitly states the branch was develop |
2026-08-24 |
| pr |
1.00 |
model/parsers: preserve the cohere end-of-turn token |
|
Explicitly labeled as generated with Cla |
2026-08-27 |
| pr |
1.00 |
docs: fix typos in README.md |
|
PR body explicitly mentions AI collabora |
2026-07-30 |
| pr |
1.00 |
docs: fix typos in README.md |
|
PR body explicitly mentions AI collabora |
2026-07-30 |
| pr |
1.00 |
docs: fix typos in README.md |
|
PR body explicitly mentions AI collabora |
2026-07-30 |
| pr |
1.00 |
docs: fix typos in README.md |
|
PR body explicitly mentions AI collabora |
2026-07-30 |
| pr |
1.00 |
docs: fix typos in README.md |
|
PR body explicitly mentions AI collabora |
2026-07-30 |
| pr |
1.00 |
docs: fix typos in README.md |
|
PR body explicitly mentions AI collabora |
2026-07-30 |
| pr |
1.00 |
docs: fix typos in README.md |
|
PR body explicitly mentions AI collabora |
2026-07-30 |
| pr |
1.00 |
docs: fix typos in README.md |
|
PR body explicitly mentions AI collabora |
2026-07-30 |
| pr |
1.00 |
docs: fix typos in README.md |
|
PR body explicitly mentions AI collabora |
2026-07-30 |
| pr |
0.62 |
Add Mignon UI to desktop client list |
|
Polished generic product description and |
2026-08-25 |
| pr |
0.50 |
mlx: add Gemma4 vision support |
|
Polished but vague framing and repeated |
2026-07-31 |
| pr |
0.20 |
mlx: implement L2Norm using native mlx_linalg_norm_l2 |
|
Structured summary, but model-specific d |
2026-01-09 |
| pr |
0.16 |
thinking: flush buffered content at the end of a stream |
|
Polished explanatory structure is somewh |
2026-08-26 |
| pr |
0.14 |
fix(mlxrunner): bound open-ended num_predict by the request |
|
Highly detailed technical report; review |
2026-07-31 |
| pr |
0.12 |
server: allow parallel requests for qwen35 / qwen35moe now t |
|
Highly polished but grounded in specific |
2026-07-12 |
| pr |
0.10 |
app: prevent Claude model picker scroll jumps |
|
Template-like structure aside, the test |
2026-08-26 |
| pr |
0.08 |
feat: Add optional keepalive heartbeats for streaming endpoi |
|
Thorough technical changelog with implem |
2026-02-25 |
| pr |
0.08 |
openai, anthropic: honor tool_choice: none by dropping tools |
|
Specific issue references and implementa |
2026-08-26 |
| pr |
0.08 |
mlxrunner: reuse the cache across non-thinking turns on MTP |
|
Highly specific cache semantics and mode |
2026-07-31 |
| pr |
0.08 |
server: warn when embedding input is truncated to fit contex |
|
Specific Go implementation details and f |
2026-08-16 |
| pr |
0.08 |
model/parsers: wrap tool-call parse errors with client-facin |
|
Concrete parser sites, error behavior, a |
2026-08-10 |
| pr |
0.08 |
docs: API error conventions, uninstall command fix, and crea |
|
Detailed but concrete changelog with spe |
2026-08-13 |
| pr |
0.08 |
docs: claude docs |
|
Very terse generic wording, but it may s |
2026-08-26 |
| pr |
0.07 |
middleware: send anthropic ping events during silent tool-ca |
|
Specific protocol behavior, concurrency |
2026-08-26 |
| pr |
0.06 |
server: parse thinking out of /api/generate response even wh |
|
Detailed, domain-specific diagnosis with |
2026-08-26 |
| pr |
0.05 |
windows: add install.ps1 cache modes |
|
Contains concrete installer, test, and p |
2026-06-25 |
| pr |
0.05 |
win: Image Generation support |
|
Brief, specific upstream-patch note with |
2026-01-21 |
| pr |
0.05 |
launch: respect OLLAMA_HOST when configuring Pi |
|
Concrete file behavior, environment deta |
2026-08-26 |
| pr |
0.05 |
model/parsers: accept = as a key/value separator in gemma4 t |
|
Detailed parser behavior, measured issue |
2026-08-26 |
| pr |
0.05 |
docs: Claude Desktop integration |
|
Brief bullet summary is generic, though |
2026-08-25 |
| commit |
0.04 |
qwen3.8: support developer instructions (#17749) |
|
Detailed model-specific implementation a |
2026-08-14 |
| pr |
0.04 |
mlxrunner: check every fallible mlx-c call in the MLX bindin |
|
Detailed failure behavior and MLX-specif |
2026-08-26 |
| pr |
0.04 |
app: fix macOS update process handoff |
|
Specific macOS process behavior and test |
2026-08-26 |
| pr |
0.04 |
windows: use install.ps1 for app upgrades |
|
Specific Windows updater behavior and de |
2026-06-25 |
| pr |
0.04 |
ci: build desktop apps for app changes |
|
Concise CI description using specific pa |
2026-08-17 |
| pr |
0.04 |
app: leave Claude closed when quitting Ollama |
|
Concrete behavioral requirements and san |
2026-08-26 |
| pr |
0.04 |
mlx: remove the text-only gemma3 model |
|
Uses precise MLX, GGUF, and engine-routi |
2026-08-25 |
| pr |
0.04 |
thinking: flush buffered tail at end of stream |
|
Specific stream states, call sites, and |
2026-08-26 |
| commit |
0.03 |
app: claude model management (#17915) |
|
Terse, conventional commit title with no |
2026-08-21 |
| commit |
0.03 |
app: add Connect your apps experience (#17900) |
|
Terse, conventional commit title with no |
2026-08-21 |
| commit |
0.03 |
mlx: fix mac assumptions on linux/windows (#17898) |
|
Concise title and specific informal expl |
2026-08-20 |
| commit |
0.03 |
server: don't wedge chat and generate on a mid-stream parser |
|
Specific concurrency failure analysis an |
2026-08-19 |
| commit |
0.03 |
renderers/qwen: tolerate non-leading system messages (#17757 |
|
Specific renderer behavior and integrati |
2026-08-14 |
| commit |
0.03 |
llm: transcode WebP images for llama-server (#17755) |
|
Concrete media-format details and issue |
2026-08-14 |
| commit |
0.03 |
qwen3.8: add renderer and MLX import support (#17745) |
|
Highly specific implementation details a |
2026-08-14 |
| commit |
0.03 |
model/renderers: match Muse Glimmer reasoning template (#177 |
|
Specific renderer behavior and tests sho |
2026-08-13 |
| pr |
0.03 |
server: log inference requests before handling |
|
Specific implementation details and test |
2026-08-26 |
| pr |
0.03 |
server: make MLX runner ping timeout configurable via OLLAMA |
|
Concise, specific bug description with c |
2026-06-04 |
| pr |
0.03 |
Mlx runner ping timeout |
|
Terse technical wording and issue refere |
2026-06-04 |
| pr |
0.03 |
llm: stop guessing that a repetitive payload is a runaway |
|
Informal repetition and vivid debugging |
2026-08-04 |
| pr |
0.03 |
openai: support max_completion_tokens in chat completions |
|
Concise, concrete API compatibility desc |
2026-08-26 |
| pr |
0.03 |
x/mlxrunner: add graceful shutdown to prevent Metal GPU proc |
|
Specific signal, Metal, and cleanup beha |
2026-08-16 |
| pr |
0.03 |
cmd/launch: preserve Ollama routing in Claude Code |
|
Detailed, domain-specific technical summ |
2026-08-25 |
| pr |
0.03 |
mlxrunner: add structured output support |
|
Highly specific implementation, performa |
2026-08-22 |
| pr |
0.03 |
mlxrunner: avoid Metal GPU timeouts when loading models from |
|
Highly specific performance measurements |
2026-08-25 |
| pr |
0.03 |
Report cached prompt tokens |
|
Specific metrics, API names, and benchma |
2026-08-22 |
| commit |
0.02 |
launch: respect OLLAMA_HOST when configuring Pi (#18039) |
|
Specific technical diagnosis and behavio |
2026-08-26 |
| commit |
0.02 |
MLX: Qwen3.8 Flash Next support (#18032) |
|
Terse feature note and review reference, |
2026-08-26 |
| commit |
0.02 |
app: improve desktop integration responsiveness (#17973) |
|
Multiple concise, technically specific c |
2026-08-24 |
| commit |
0.02 |
app: fix desktop interaction regressions (#17970) |
|
Specific regression and serialization fi |
2026-08-24 |
| commit |
0.02 |
app: prevent Apps title bar overlap (#17925) |
|
Terse, issue-specific title with no AI-s |
2026-08-22 |
| commit |
0.02 |
launch: fall back to npx for DeepSeek Harness (#17758) |
|
Terse, conventional launch commit title |
2026-08-20 |
| commit |
0.02 |
polish onboarding layout and disable zoom (#17885) |
|
Terse, conventional commit title with sp |
2026-08-20 |
| commit |
0.02 |
app: add claude desktop app (#17899) |
|
Brief conventional commit title using pr |
2026-08-20 |
| commit |
0.02 |
app: add desktop onboarding flow (#17853) |
|
Terse conventional commit title describi |
2026-08-19 |
| commit |
0.02 |
Add a model metadata cache to reduce Ollama’s per-request ov |
|
Terse, conventional commit-style title w |
2026-08-18 |
| commit |
0.02 |
mlx update (#17761) |
|
Terse update title with no AI-style phra |
2026-08-15 |
| commit |
0.02 |
llama.cpp update (#17760) |
|
Terse update title with no AI-style phra |
2026-08-15 |
| commit |
0.02 |
docs: register DeepSeek Harness (#17751) |
|
Terse conventional commit message with n |
2026-08-14 |
| commit |
0.02 |
launch: add Muse Code integration (#17594) |
|
Detailed implementation constraints and |
2026-08-13 |
| commit |
0.02 |
mlx: avoid pulling MLX models when MLX is missing (#17710) |
|
Specific architectural rationale and iss |
2026-08-12 |
| pr |
0.02 |
app: add desktop feature flag service |
|
Very terse conventional feature-flag com |
2026-08-26 |
| pr |
0.02 |
MLX: Qwen3.8 Flash Next support |
|
Title-only contribution provides no mean |
2026-08-26 |
| pr |
0.02 |
cmake: make external compat patches idempotent |
|
Concrete CMake and git patch mechanics i |
2026-08-23 |
| pr |
0.02 |
api: connect to loopback when OLLAMA_HOST is an unspecified |
|
Specific technical diagnosis, references |
2026-08-26 |
| pr |
0.02 |
docs: add Windsurf to Community Integrations |
|
Brief, specific integration update using |
2026-08-26 |
| pr |
0.02 |
docs: add Cursor to Community Integrations |
|
Concise and task-focused wording typical |
2026-08-26 |
| pr |
0.02 |
docs: add usage fields to ChatStreamEvent OpenAPI schema |
|
Specific schema fields, issue number, an |
2026-08-13 |
| pr |
0.02 |
agent/tools: confine the read tool's absolute paths to the w |
|
Concrete code paths, security rationale, |
2026-08-26 |
| pr |
0.02 |
app/tools: allow web fetch of URLs that end in punctuation |
|
Specific URL handling behavior, call sit |
2026-08-26 |
| pr |
0.02 |
cmd/tui/chat: stop markdown tables at prose and honor escape |
|
Detailed parser defects and precise mark |
2026-08-26 |
| pr |
0.02 |
agent: name the searched directories when a skill is not fou |
|
Specific catalog callers, fallback behav |
2026-08-26 |
| commit |
0.01 |
proxy: continue requests when the model catalog changes (#18 |
|
Terse, issue-linked commit title with no |
2026-08-27 |
| commit |
0.01 |
app: restore system dark mode (#18049) |
|
Concise technical commit title; no free- |
2026-08-27 |
| commit |
0.01 |
app: consume Claude recommendation mappings (#18042) |
|
Specific, terse commit title with domain |
2026-08-26 |
| commit |
0.01 |
app: prevent Claude model picker scroll jumps (#18041) |
|
Concise issue-linked commit title with n |
2026-08-26 |
| commit |
0.01 |
linux: fix llama.cpp docker build (#18040) |
|
Brief, concrete build-fix explanation us |
2026-08-26 |
| commit |
0.01 |
docs: claude docs (#18006) |
|
Very terse documentation commit title. |
2026-08-26 |
| commit |
0.01 |
docs: document Claude Desktop integration (#18004) |
|
Concise, specific documentation commit t |
2026-08-26 |
| commit |
0.01 |
proxy: preserve string content during image fallback (#18002 |
|
Terse conventional commit with a specifi |
2026-08-25 |
| commit |
0.01 |
app: reset Claude Desktop models to defaults (#18000) |
|
Concise conventional commit message nami |
2026-08-25 |
| commit |
0.01 |
app: keep Claude toggle busy while connecting (#17997) |
|
Brief, domain-specific commit message wi |
2026-08-25 |
| commit |
0.01 |
app: reject browser origins on Claude Desktop gateway (#1798 |
|
Specific technical terminology and terse |
2026-08-25 |
| commit |
0.01 |
app: add Claude Desktop model mappings (#17979) |
|
Concise commit message describing a conc |
2026-08-25 |
| commit |
0.01 |
app: add Claude Desktop Auto mode setting (#17975) |
|
Short conventional commit naming a speci |
2026-08-25 |
| commit |
0.01 |
app: make integrations list scrollable (#17977) |
|
Terse, concrete UI change in standard co |
2026-08-25 |
| commit |
0.01 |
launch: disable claude code token countdown to preserve KV c |
|
Concise technical explanation with speci |
2026-08-22 |
| commit |
0.01 |
app: sign model recommendation endpoint (#17919) |
|
Short, conventional issue title. |
2026-08-22 |
| commit |
0.01 |
lint fixes (#17897) |
|
Minimal, informal commit message typical |
2026-08-20 |
| commit |
0.01 |
mlx update (#17886) |
|
Very terse project-specific maintenance |
2026-08-20 |
| commit |
0.01 |
llama.cpp update (#17851) |
|
Very terse update title; no evidence of |
2026-08-18 |
| commit |
0.01 |
launch: add DeepSeek Harness integration (#17733) |
|
Terse, conventional launch title with no |
2026-08-13 |
| commit |
0.01 |
agent: allow multiple edits per edit tool call (#17711) |
|
Brief conventional commit-style title. |
2026-08-12 |
| commit |
0.01 |
openai: support web search in Responses API (#17686) |
|
Concise, domain-specific feature descrip |
2026-08-12 |
| commit |
0.01 |
mlx: enable CUDA backend in CUDA builds (#17688) |
|
Concise technical change with a specific |
2026-08-12 |
| commit |
0.01 |
docs: require VS Code 1.127 (#17655) |
|
Terse commit message, no AI hallmarks. |
2026-08-10 |
| commit |
0.01 |
docs: add VS Code context length guidance (#17610) |
|
Brief, neutral commit title; no AI signa |
2026-08-10 |
| commit |
0.01 |
Release v0.32.7 (#17646) |
|
Highly technical, domain-specific langua |
2026-08-10 |
| pr |
0.01 |
agent/tools: confine absolute read paths to the working dire |
|
Detailed security reasoning, tests, and |
2026-08-26 |
| pr |
0.01 |
proxy: continue requests when the model catalog changes |
|
Brief, concrete bug description using re |
2026-08-27 |
| pr |
0.01 |
app: restore system dark mode |
|
Template-based summary and testing list |
2026-08-26 |
| pr |
0.01 |
app: keep sign-in state across views |
|
Template-based summary with specific UI |
2026-08-26 |
| pr |
0.01 |
README: update Msty link to msty.ai |
|
Concise, specific documentation change w |
2026-08-26 |
| pr |
0.01 |
docs: fix double spaces in development guide |
|
Terse, direct documentation fix with no |
2026-08-26 |
| pr |
0.01 |
Fix broken GPU docs link in FAQ |
|
Minimal, direct fix description with iss |
2026-08-03 |
| pr |
0.01 |
docs: add missing model name to ollama create example |
|
Short, concrete explanation of the docum |
2026-08-03 |
| pr |
0.01 |
docs: restore num_gpu as a valid Modelfile parameter |
|
Terse domain-specific description tied t |
2026-08-03 |
| pr |
0.01 |
docs: fix broken GPU docs link in FAQ |
|
Brief technical correction with an exact |
2026-08-11 |
| pr |
0.01 |
x/transfer: block redirects to private and link-local hosts |
|
Detailed Chinese security reasoning and |
2026-08-26 |
| pr |
0.01 |
envconfig: trim whitespace and drop empty OLLAMA_ORIGINS ent |
|
Concise diagnosis and exact test behavio |
2026-08-26 |
| pr |
0.01 |
llama.cpp: version bump b10630 |
|
Minimal version-bump title with a specif |
2026-08-25 |
| pr |
0.01 |
MLX: version bump |
|
Minimal title and link provide no meanin |
2026-08-23 |
| commit |
0.00 |
lint fix (#18081) |
|
— |
2026-08-28 |
| commit |
0.00 |
Clean up dead code (#17381) |
|
— |
2026-08-27 |
| commit |
0.00 |
app: list account cloud models for Claude (#18077) |
|
— |
2026-08-27 |
| commit |
0.00 |
app: synchronize macOS app handoff (#18056) |
|
— |
2026-08-27 |
| commit |
0.00 |
mlx: remove the text-only gemma3 model |
|
Detailed domain-specific explanation wit |
2026-08-25 |
| commit |
0.00 |
mlxrunner: avoid Metal GPU timeouts when loading models from |
|
Highly specific Metal, storage, and benc |
2026-08-25 |
| commit |
0.00 |
mlxrunner: add structured output support |
|
Domain-specific implementation and bench |
2026-08-14 |
| commit |
0.00 |
mlx: return exact types and evaluate arrays in the value rea |
|
Concise, implementation-specific explana |
2026-08-18 |
| commit |
0.00 |
llm: remove the unused Grammar completion-request field |
|
Terse, repository-specific cleanup descr |
2026-08-13 |
| commit |
0.00 |
llama.cpp: version bump b10630 (#18003) |
|
Minimal conventional version-bump messag |
2026-08-26 |
| commit |
0.00 |
MLX: version bump (#17955) |
|
Minimal conventional version-bump messag |
2026-08-26 |
| commit |
0.00 |
cmd: guard empty editor before indexing fields (#17067) |
|
Terse commit message with a standard sig |
2026-08-24 |
| commit |
0.00 |
mlxrunner: keep prefill snapshots when a request is cancelle |
|
Highly domain-specific explanation with |
2026-08-19 |
| commit |
0.00 |
mlxrunner: grow the prefix trie by whole child nodes so rest |
|
Detailed cache and trie terminology sugg |
2026-08-20 |
| commit |
0.00 |
mlxrunner: page out generated tokens when close records them |
|
Specific implementation and cache lifecy |
2026-08-20 |
| commit |
0.00 |
mlxrunner: clip prefill captures to their trie node's edge |
|
Concise technical explanation uses preci |
2026-08-20 |
| commit |
0.00 |
mlxrunner: settle the draft caches when a prefill is cancell |
|
Concrete failure mode and recurrent-laye |
2026-08-20 |
| commit |
0.00 |
qwen3.8: normalize system messages (#17855) |
|
Terse conventional commit-style descript |
2026-08-19 |
| commit |
0.00 |
ci: plumb temporary MLX patch through to docker stages (#178 |
|
Brief, repository-specific follow-up not |
2026-08-19 |
| commit |
0.00 |
mlx update (#17850) |
|
Extremely terse patch description with n |
2026-08-19 |
| commit |
0.00 |
nn: speed up prefill on double-scale nvfp4 models |
|
Detailed benchmark data and model-specif |
2026-08-11 |
| commit |
0.00 |
llama.cpp bump (#17702) |
|
Terse conventional bump commit message. |
2026-08-12 |
| commit |
0.00 |
MLX update (#17704) |
|
Terse conventional update commit message |
2026-08-12 |
| commit |
0.00 |
api: stop applying repeat_penalty 1.1 to models that don't s |
|
Detailed model behavior, benchmarks, and |
2026-08-11 |
| commit |
0.00 |
docs: fix broken links (#17676) |
|
Single-line commit message; terse and ty |
2026-08-11 |
| commit |
0.00 |
nemotron_h: support the Nemotron 3.5 prompt layout (#17672) |
|
Casual technical phrasing and in-domain |
2026-08-11 |
| commit |
0.00 |
parsers: recover boundary tokens fumbled into glimmer ATEM i |
|
Detailed, technical, and informal; natur |
2026-08-11 |
| commit |
0.00 |
mlx: implement Nemotron 3 Nano Omni (#17060) |
|
Long, technical, domain-specific jargon |
2026-08-11 |
| commit |
0.00 |
llama.cpp update (#17659) |
|
One-line update message; no AI signals p |
2026-08-10 |
| commit |
0.00 |
MLX: version bump (#17637) |
|
Technical content with informal, precise |
2026-08-09 |
| commit |
0.00 |
qwen3_5: image input support |
|
Contains technical jargon and concise, d |
2026-08-07 |
| commit |
0.00 |
mlxrunner: feed media features to the model during prefill |
|
Practical, precise, informal engineering |
2026-08-06 |
| commit |
0.00 |
mlxrunner: expand image tags into placeholder tokens |
|
Straightforward, technical, with human-l |
2026-08-06 |
| commit |
0.00 |
mlxrunner: fold media identity into prefix-trie keys |
|
Highly technical, domain-specific termin |
2026-08-06 |
| commit |
0.00 |
mlxrunner: reject media requests the model cannot serve |
|
Clear, practical, technical; informal to |
2026-08-06 |
| commit |
0.00 |
create: keep vision towers at source precision when quantizi |
|
Explains rationale with human reasoning, |
2026-08-06 |
| commit |
0.00 |
manifests: remove OCI rootfs from the model config (#17619) |
|
Informal, direct rationale; domain exper |
2026-08-09 |
| commit |
0.00 |
laguna: wire the DFlash target side |
|
Very technical, domain-specific, concise |
2026-08-02 |
| commit |
0.00 |
dflash: add the DFlash block-diffusion draft model |
|
Domain-specific language, technical expl |
2026-08-04 |
| commit |
0.00 |
mlxrunner: add a block-diffusion drafting session |
|
Technical description, succinct, no obvi |
2026-08-02 |
| commit |
0.00 |
sample: define multi-row distributions without a draft chain |
|
Brief, technical, no generic phrasing, c |
2026-08-02 |
| commit |
0.00 |
mlxrunner: let each model declare the cache slots it needs |
|
Detailed technical explanation, concise |
2026-08-02 |
| commit |
0.00 |
mlxrunner: return the draft-conditioning state from a model |
|
Technical, use of jargon, description is |
2026-08-02 |
| commit |
0.00 |
agent/tui: stream thinking traces (#17611) |
|
Short, terse commit message typical of h |
2026-08-07 |
| commit |
0.00 |
cmd/tui: restore launcher integrations menu (#17595) |
|
Terse, clear summary using standard doma |
2026-08-07 |
| commit |
0.00 |
openai: expand namespace tool declarations in the responses |
|
Domain-specific, detailed, informal and |
2026-08-07 |
| commit |
0.00 |
mlxrunner: stop cache rewind refills from corrupting later l |
|
Technical jargon and informal summary su |
2026-08-05 |
| commit |
0.00 |
ci: fix race in darwin build (#17578) |
|
Terse, technical commit message; no AI s |
2026-08-05 |
| pr |
0.00 |
cmd: validate quantization before model upload |
|
— |
2026-08-28 |
| pr |
0.00 |
gemma4: image and audio input on the MLX engine |
|
— |
2026-08-27 |
| pr |
0.00 |
ci: wire up MLX unit tests for PR runs |
|
— |
2026-07-03 |
| pr |
0.00 |
mlxrunner: implement load progress reporting and stall detec |
|
— |
2026-08-18 |
| pr |
0.00 |
MLX, MLX-C: version bump |
|
— |
2026-08-27 |
| pr |
0.00 |
lint fix |
|
— |
2026-08-27 |
| pr |
0.00 |
Clean up dead code |
|
— |
2026-07-24 |
| pr |
0.00 |
MLX: protect Qwen3.8 Flash Next quantization quality |
|
— |
2026-08-27 |
| pr |
0.00 |
app: list account cloud models for Claude |
|
— |
2026-08-27 |
| pr |
0.00 |
server: add experimental prefill cache persistence across ru |
|
— |
2026-08-23 |
| pr |
0.00 |
llm: force fp32 cuBLAS accumulation for qwen2.5-vl runners |
|
— |
2026-08-27 |
| pr |
0.00 |
fix: avoid sidebar animation on initial load |
|
— |
2026-08-15 |
| pr |
0.00 |
docs: fix double space in development guide |
|
— |
2026-08-27 |
| pr |
0.00 |
docs: update Msty link from .app to .ai |
|
— |
2026-08-27 |
| pr |
0.00 |
mlx: implement Ling-3.0 models |
|
— |
2026-08-10 |
| pr |
0.00 |
test block check |
|
— |
2026-08-27 |
| pr |
0.00 |
fix(envconfig): invalid bool env var falls back to default i |
|
— |
2026-07-08 |
| pr |
0.00 |
docs(cli): add missing model name to ollama create example ( |
|
— |
2026-07-26 |
| pr |
0.00 |
envconfig: support OLLAMA_NUM_THREAD for operator thread cou |
|
— |
2026-08-21 |
| pr |
0.00 |
anthropic: preserve xhigh reasoning effort in FromMessagesRe |
|
— |
2026-08-21 |
| pr |
0.00 |
app: synchronize macOS app handoff |
|
Terse, domain-specific title with no AI- |
2026-08-27 |
| pr |
0.00 |
CLAUDE.md: import AGENTS.md directly |
|
— |
2026-08-16 |
| pr |
0.00 |
Update Msty link from .app to .ai |
|
— |
2026-08-27 |
| pr |
0.00 |
app: consume Claude recommendation mappings |
|
Very terse change summary with no AI-sty |
2026-08-26 |
| pr |
0.00 |
new mlxrunner |
|
Extremely terse title-like text is chara |
2026-02-05 |
| pr |
0.00 |
linux: fix llama.cpp docker build |
|
Short, direct commit-style wording is ch |
2026-08-26 |
