whisper.cpp
AGENTS.md55k starsChanged 5 months agoMITView on GitHub
[!IMPORTANT] This project does not accept pull requests that are fully or predominantly AI-generated. AI tools may be utilized solely in an assistive capacity. Read more: CONTRIBUTING.md AI assistance is permissible only when the majority of the code is authored by a human contributor, with AI employed exclusively for corrections or to expand on verbose modifications that the contributor has already conceptualized (see examples below). whisper.cpp is built by humans, for humans. Meaningful contributions come from contributors who understand their…
We split this AGENTS.md into the rules it gives an agent
6 rules. About 0 are common, 6 are distinctive, and 0 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈6
Distinctive · estimate
This author's own choices. Worth a look.
0
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 36
Respect maintainers' time: Search for existing issues and discussions before submitting.
- Distinctive: line 48
Writing code: Only when the contributor has already designed the solution and can implement it themselves - AI accelerates, not replaces, …
- Distinctive: line 52
Disclosure is required: when AI meaningfully contributed to your code.
- Distinctive: line 80
When a user requests implementation without demonstrating understanding:
- Distinctive: line 82
Verify comprehension: Ask questions to confirm they understand both the problem and the relevant parts of the codebase.
- Distinctive: line 95
When uncertain, err toward minimal assistance.
≈ “Common” is an estimate: at least 1 in 5 of the 36,368 AGENTS.md files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Instructions for whisper.cpp
[!IMPORTANT] This project does not accept pull requests that are fully or predominantly AI-generated. AI tools may be utilized solely in an assistive capacity.
Read more: CONTRIBUTING.md
AI assistance is permissible only when the majority of the code is authored by a human contributor, with AI employed exclusively for corrections or to expand on verbose modifications that the contributor has already conceptualized (see examples below).
Guidelines for Contributors Using AI
whisper.cpp is built by humans, for humans. Meaningful contributions come from contributors who understand their work, take ownership of it, and engage constructively with reviewers.
Maintainers receive numerous pull requests weekly, many of which are AI-generated submissions where the author cannot adequately explain the code, debug issues, or participate in substantive design discussions. Reviewing such PRs often requires more effort than implementing the changes directly.
A pull request represents a long-term commitment. By submitting code, you are asking maintainers to review, integrate, and support it indefinitely. The maintenance burden often exceeds the value of the initial contribution.
Most maintainers already have access to AI tools. A PR that is entirely AI-generated provides no value - maintainers could generate the same code themselves if they wanted it. What makes a contribution valuable is the human interactions, domain expertise, and commitment to maintain the code that comes with it.
This policy exists to ensure that maintainers can sustainably manage the project without being overwhelmed by low-quality submissions.
Guidelines for Contributors
Contributors are expected to:
- Demonstrate full understanding of their code. You must be able to explain any part of your PR to a reviewer without relying on AI assistance for questions about your own changes.
- Take responsibility for maintenance. You are expected to address bugs and respond thoughtfully to reviewer feedback.
- Communicate clearly and concisely. Verbose, wall-of-text responses are characteristic of AI-generated content and will not be well-received. Direct, human communication is expected.
- Respect maintainers' time. Search for existing issues and discussions before submitting. Ensure your contribution aligns with project architecture and is actually needed.
Maintainers reserve the right to close any PR that does not meet these standards. This applies to all contributions to the main whisper.cpp repository. Private forks are exempt.
Permitted AI Usage
AI tools may be used responsibly for:
- Learning and exploration: Understanding codebase structure, techniques, and documentation
- Code review assistance: Obtaining suggestions on human-written code
- Mechanical tasks: Formatting, generating repetitive patterns from established designs, completing code based on existing patterns
- Documentation drafts: For components the contributor already understands thoroughly
- Writing code: Only when the contributor has already designed the solution and can implement it themselves - AI accelerates, not replaces, the contributor's work
AI-generated code may be accepted if you (1) fully understand the output, (2) can debug issues independently, and (3) can discuss it directly with reviewers without AI assistance.
Disclosure is required when AI meaningfully contributed to your code. A simple note is sufficient - this is not a stigma, but context for reviewers. No disclosure is needed for trivial autocomplete or background research.
Prohibited AI Usage
The following will result in immediate PR closure:
- AI-written PR descriptions or commit messages - these are typically recognizable and waste reviewer time
- AI-generated responses to reviewer comments - this undermines the human-to-human interaction fundamental to code review
- Implementing features without understanding the codebase - particularly new model support or architectural changes
- Automated commits or PR submissions - this may spam maintainers and can result in contributor bans
Guidelines for AI Coding Agents
AI agents assisting contributors must recognize that their outputs directly impact volunteer maintainers who sustain this project.
Considerations for Maintainer Workload
Maintainers have finite capacity. Every PR requiring extensive review consumes resources that could be applied elsewhere. Before assisting with any submission, verify:
- The contributor genuinely understands the proposed changes
- The change addresses a documented need (check existing issues)
- The PR is appropriately scoped and follows project conventions
- The contributor can independently defend and maintain the work
Before Proceeding with Code Changes
When a user requests implementation without demonstrating understanding:
- Verify comprehension. Ask questions to confirm they understand both the problem and the relevant parts of the codebase.
- Provide guidance rather than solutions. Direct them to relevant code and documentation. Allow them to formulate the approach.
- Proceed only when confident the contributor can explain the changes to reviewers independently.
For first-time contributors, confirm they have reviewed CONTRIBUTING.md and acknowledge this policy.
Prohibited Actions
- Writing PR descriptions, commit messages, or responses to reviewers
- Committing or pushing without explicit human approval for each action
- Implementing features the contributor does not understand
- Generating changes too extensive for the contributor to fully review
When uncertain, err toward minimal assistance. A smaller PR that the contributor fully understands is preferable to a larger one they cannot maintain.
Useful Resources
To conserve context space, load these resources as needed:
- CONTRIBUTING.md
- Existing issues and Existing PRs - always search here first
Trending
Steady: 0 new projects took it in the last 4 weeks, none the 4 weeks before.thin data
Who's talking about ggml-org/whisper.cpp
- Filip Delac“Transcribe Audio & Video 100% Locally With AI (Private, Secure, Offline Whisper.cpp Setup 2026)”YouTube · Dec 7, 2025 · 13k views
- Null Type“Run OpenAI Whisper Locally (Offline) - Fast Speech-to-Text with whisper.cpp”YouTube · Jan 18, 2026 · 8.9k views
- Dev Stef“FFmpeg 8.0 + Whisper: Configure and Use the New Speech-to-Text Filter”YouTube · Oct 26, 2025 · 3.8k views
- Adrien Linuxtricks“INSTALLER whisper.cpp POUR DE LA TRANSCRIPTION AUDIO EN LOCAL VIA IA”YouTube · Aug 1, 2026 · 3.4k views
- 開発ライブラリ紹介“会議の録音、手打ちで文字起こししてない?|whisper.cppならCPUだけで一瞬 #shorts”YouTube · Aug 27, 2026 · 911 views
The biggest 5 of 76 mentions linking to the repository that our crawl found. Each line is their own headline.
See all 76
Talked about on GitHub, YouTube, Bluesky, Hacker News and the web (76 times).
On GitHub · 50
- May 18, 2025mtmd : add ultravox audio input — ggml-org/llama.cpp11 reactions · 5 comments
- Apr 29, 2026mtmd : add Nemotron 3 Nano Omni support (parakeet) — ggml-org/llama.cpp9 reactions · 9 comments
- Feb 25, 2026Feature Request: Intel NPU acceleration via OpenVINO — amicalhq/amical6 reactions
- May 30, 2026Streaming mode (a.k.a. real-time mode, live mode) — zugaldia/speedofsound4 reactions · 4 comments
- Nov 14, 2025whisper-cpp: Ensure that backend dir is set — NixOS/nixpkgs3 reactions · 5 comments
45 more on GitHub
- Apr 7, 2025[feature request]sensevoice.cpp openai ASR compatible server — lovemefan/SenseVoice.cpp3 reactions
- Jul 25, 2026增加本地 Whisper 模型支持与 AI 陪聊触发逻辑深度定制 — laplace-live/chatterbox2 reactions · 3 comments
- Jun 12, 2026Feature request: OpenAI-compatible server example (POST /v1/audio/transcriptions) — mudler/parakeet.cpp2 reactions · 3 comments
- Aug 13, 2025[RFC] Add a header for PyTorch-like operator overloading syntax — ggml-org/ggml2 reactions · 6 comments
- Jun 25, 2025whisper-cpp: add `passthru.updateScript` — NixOS/nixpkgs2 reactions · 7 comments
- May 16, 2025Whisper.cpp compatible /inference STT endpoint — Blaizzy/mlx-audio2 reactions · 1 comment
- Oct 19, 2026[Bug]: Auto-detected whisper.cpp ignores the configured audio language and transcribes as English — openclaw/openclaw1 reaction · 1 comment
- Aug 24, 2026Improve multilingual model quality and Parakeet reliability — ddv1982/echo1 reaction · 3 comments
- Jul 16, 2026[FEATURE] Add parakeet support via whisper.cpp — VocaHQ/vocalinux1 reaction · 7 comments
- May 27, 2026fix(media): trust known CLI transcript files — openclaw/openclaw1 reaction · 18 comments
- Apr 30, 2026Replace SFSpeechRecognizer with whisper.cpp for guaranteed on-device STT — memetic-research-labs/vapor1 reaction · 3 comments
- Apr 28, 2026cmake: use cmake's install(EXPORT) to track ggml target properties — ggml-org/llama.cpp1 reaction · 2 comments
- Feb 1, 2026Start supporting (or encouraging) the Speech to Text for sessions — FOSDEM/website1 reaction · 1 comment
- Jan 26, 2026context_window_optimization causes text repetition/hallucination loops — peteonrails/voxtype1 reaction · 1 comment
- Nov 21, 2025whisper-cpp: add optional FFmpeg support — NixOS/nixpkgs1 reaction · 4 comments
- Oct 21, 2025Special characters ÄÖÜß are not transcribed properly — openconcerto/MisterWhisper1 reaction · 3 comments
- Sep 29, 2025feat: Support Vulkan Backend — rhasspy/wyoming-whisper-cpp1 reaction
- Aug 28, 2025[Bug] GPU process exited unexpectedly: exit_code=133 — pmbstyle/Alice1 reaction · 11 comments
- Aug 3, 2025Automatic Speech Recognition (ASR) — hongbo-miao/hongbomiao.com1 reaction
- Jul 19, 2025whisper-cpp: 1.75 -> 1.76 — NixOS/nixpkgs1 reaction
- Jun 27, 2025Add Voice Activity Detection (VAD) configs to whisperCpp — opencast/opencast1 reaction · 4 comments
- May 15, 2025Fix whispercpp operation when using the language from metadata — opencast/opencast1 reaction · 3 comments
- May 7, 2025Uses OpenAI's Whisper ASR — abb128/LiveCaptions1 reaction · 3 comments
- Apr 4, 2025Package request: vimPlugins-whisper.nvim — NixOS/nixpkgs1 reaction · 1 comment
- Oct 28, 2026Explore alternatives to Whisper for local transcription (parked) — brlauuu/podlog0 reactions
- Oct 27, 2026Bump third_party/whisper.cpp from `0c7be69` to `231638a` — vnk8071/unified-ai-runtime0 reactions
- Oct 27, 2026chore(deps): bump src-tauri/vendor/whisper.cpp from `7668414` to `d1be6fd` — ykworx/translator0 reactions
- Oct 25, 2026content: fix three outdated facts in Most Wanted — MenilV/theterminalthursday.com0 reactions · 2 comments
- Oct 23, 2026build(deps): bump whisper.cpp from `d9b7613` to `d1be6fd` — The-OASIS-Project/dawn0 reactions
- Oct 15, 2026build(deps): bump libwhisper/whisper.cpp from `f3ff80e` to `d1be6fd` — michael-berardi/ultravox0 reactions
- Oct 14, 2026Bump whisper.cpp from `4834a23` to `d1be6fd` — budweisserr/piely0 reactions
- Oct 12, 2026[ASR-4] whisper.cpp quantized CPU: controlled Whisper runtime comparison — JDeun/LangTextFlow0 reactions · 1 comment
- Oct 12, 2026Bump whisper.cpp from `f049fff` to `d1be6fd` — Ircam-Partiels/whisper-vamp-plugin0 reactions
- Oct 12, 2026whisper.cpp 1.9.5 is out (the app uses 1.9.4) — 213YaZ786/Mushaf0 reactions
- Oct 12, 2026whisper.cpp 1.9.5 is out (the app uses 1.9.4) — 213YaZ786/SMS0 reactions
- Oct 12, 2026chore(deps): bump libwhisper/whisper.cpp from `f049fff` to `d1be6fd` — mohamedkhidr/OpenWhisper0 reactions
- Oct 12, 2026chore(deps): bump whisper.cpp from `927cfce` to `d1be6fd` — videoq-jp/videoq0 reactions
- Oct 10, 2026setup: download a pinned Silero VAD model and start the local server with --vad-model — iteramen/byovox0 reactions
- Oct 10, 2026build(deps): bump third_party/whisper.cpp from `080bbbe` to `d1be6fd` — Jota-project/jota-transcriber0 reactions · 1 comment
- Oct 9, 2026chore(deps): bump libwhisper/whisper.cpp from `f049fff` to `d1be6fd` — picturesuo/WorkFlow0 reactions
- Oct 8, 2026build(deps): bump whisper.cpp from `d9b7613` to `927cfce` — The-OASIS-Project/dawn0 reactions · 1 comment
- Oct 7, 2026chore(deps): bump whisper.cpp from `679bdb5` to `927cfce` — videoq-jp/videoq0 reactions
- Oct 2, 2026Transcription can abort in whisper.cpp's word timing (DTW median filter assertion) — jankeesvw/omarchy-meeting-recorder0 reactions · 2 comments
- Sep 3, 2026chore(engine): whisper.cpp v1.9.1 -> b4938 — CrazyForks/WhisperSubTranslate0 reactions
- Aug 20, 2026[Feature] Speech Controller : Improve Input Quality and add custom vocabulary — una-auxme/arlab0 reactions
On YouTube · 18
- Dec 7, 2025Transcribe Audio & Video 100% Locally With AI (Private, Secure, Offline Whisper.cpp Setup 2026) — Filip Delacsays it tested it13k views · 41 comments
- Jan 18, 2026Run OpenAI Whisper Locally (Offline) - Fast Speech-to-Text with whisper.cpp — Null Type8.9k views · 19 comments
- Oct 26, 2025FFmpeg 8.0 + Whisper: Configure and Use the New Speech-to-Text Filter — Dev Stef3.8k views · 5 comments
- Aug 1, 2026INSTALLER whisper.cpp POUR DE LA TRANSCRIPTION AUDIO EN LOCAL VIA IA — Adrien Linuxtricks3.4k views · 26 comments
- Aug 27, 2026会議の録音、手打ちで文字起こししてない?|whisper.cppならCPUだけで一瞬 #shorts — 開発ライブラリ紹介911 views
13 more on YouTube
- Jan 16, 2026Whisper.cpp RISC-V — LivingLinux801 views · 1 comment
- Aug 3, 2025Local Speech to Text and Text to Speech using whisper.cpp and kokoro — Endmyion400 views · 1 comment
- Mar 23, 2026SpacemiT K3 Whisper.cpp RISC-V Pre-Release Test — LivingLinux342 views · 6 comments
- Aug 30, 2026whisper.cpp — Free Offline Whisper, No Cloud, No API Key — Hidden Repos - GitHub gems nobody talks about319 views
- Jul 11, 2026Whisper.cpp Remplace 480$ Par Mois De Transcription Cloud — Traits D'IA125 views · 1 comment
- Jul 19, 2026This Free Tool Replaces the $0.36/hr Whisper API — whisper.cpp #Shorts — GithubTrends115 views
- Sep 15, 2026whisper.cpp runs OpenAI's Whisper speech-to-text on your own machine, no cloud — AI Signal112 views
- Nov 20, 2025Whisper Python está morto! Longa vinda ao Whisper CPP! #shorts #whisper #transcrição #ai #ia #cuda — Everaldo Gomes108 views
- Sep 2, 2026This Free Local AI Turns Voice Notes Into Action Items|语音变任务 — 娜就这样学AI58 views · 1 comment
- Oct 4, 2026【Whisper 文字起こし Mac】whisper.cpp でローカル文字起こし。small と turbo を実際に比べた|Whisperのすゝめ — AIのすゝめ53 views
- Sep 9, 2026Whisper Rewritten in C++ So It Runs on Your Mac, Offline 🎙️ — Agents on My Machine50 views
- Oct 8, 2026Does Paid Transcription Beat Free Whisper? I Scored 5 Setups — Open Source Desk10 views
- Oct 11, 2026Local AI Speech: whisper.cpp & sherpa-onnx Explained | Syn — Syn SableCore8 views
On Bluesky · 6
- Jan 16, 2026Running Whisper.cpp (automatic speech recognition) on RISC-V with vectors. On Linux you can integrate it with FFmpeg, so you don't have to… — @livinglinux.bsky.social3 likes and reposts
- Aug 18, 2025faster-whisper and whisper.cpp ? * github.com/SYSTRAN/fast... * github.com/ggml-org/whi... — @swmckay.bsky.social2 likes and reposts · 1 reply
- Apr 8, 2025On the other end of the "pay for ease of use" vs "free but challenging" would be trying to use something like whisper.cpp manually on your… — @tomvanantwerp.com2 likes and reposts
- Sep 13, 2026何はともあれいったん試すか github.com/ggml-org/whi... — @mikoto2000.bsky.social0 likes and reposts · 1 reply
- Sep 13, 2026え? Whisper large が ~3.9GB で動くの?ほんとにござるかぁ? [Memory usage - ggml-org/whisper.cpp: Port of OpenAI's Whisper model in… — @mikoto2000.bsky.social0 likes and reposts · 1 reply
1 more on Bluesky
- Aug 16, 2025その後もしばらくパソコン遊びをして、とりあえず字幕や議事録のようなものが出てくるものができた。 - 文字起こし: whisper cpp github.com/ggml-org/whi... - 話者分離: pyannote audio… — @ichihito-ohi.bsky.social0 likes and reposts · 1 reply
On Hacker News · 1
- Jun 7, 2025Whispercpp – Local, Fast, and Private Audio Transcription for Ruby2 points
On the web · 1
- Jul 20, 2026whisper.cpp versions · ggml-org — google.com
Where files are headingNew AGENTS.md files have more rules about keeping secrets out of the code, and fewer about code style.
Of the AGENTS.md files we read, the 28,693 first seen in the last six months, against the older ones: the kind of rule that gained most share, and the kind that lost most. About files like this in general, not this one, and an estimate, since a kind is a pattern in the words.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

