{"slug": "whisper", "name": "Whisper", "category": "ai", "category_name": "AI models and assistants", "description": "OpenAI's open-source speech recognition model for multilingual transcription and translation.", "recommenders": 16, "commenters": 40, "warners": 4, "score": 15.612, "rising": -1.25, "first": "2025-11-09", "latest": "2026-10-21", "why": "Commenters like it for running speech-to-text locally with accurate results, especially for dictation and multilingual transcription.", "why_is_quote": false, "has_verdict": true, "sentiment": {"recommend": 15, "neutral": 19, "compare": 3, "warn": 4}, "spark": [3, 0, 4, 1, 0, 0, 2, 1, 1, 0, 2, 1], "listed": true, "loved": ["Runs locally", "Accuracy"], "caveat": "Transcription errors", "short_quote": "My favourite use is Whisper. I voice->text almost all of my typing now.", "icon": true, "domain": null, "github_repo": null, "official_url": "https://github.com/openai/whisper", "pricing": "open source", "aliases": ["Whisper"], "verdict": [{"text": "Commenters like it for running speech-to-text locally with accurate results", "cites": [{"id": 751940768, "item_id": 46996298, "url": "https://news.ycombinator.com/item?id=46996298", "by": "srinifromsalem", "date": "2026-02-12", "stance": "recommend", "quote": "Whisper remains the gold standard - you can run it locally with good accuracy across languages.", "excerpt": "For local speech-to-text, Whisper remains the gold standard - you can run it locally with good accuracy across languages. For speech-to-speech, you'd typically chain Whisper with a local TTS model like Coqui TTS or use something like Tortoise TTS for higher quality but slower processing. The key is balancing accuracy, speed, and resource usage based on your specific use case. If you're doing content creation workflow", "context": "", "use_case": "local speech-to-text", "switched_from": "", "story": "Ask HN: What's the current best local/open speech-to-speech setup?"}, {"id": 744520368, "item_id": 46532523, "url": "https://news.ycombinator.com/item?id=46532523", "by": "delaminator", "date": "2026-01-07", "stance": "recommend", "quote": "Whisper works great, even the medium model is pretty good.", "excerpt": "Whisper works great, even the medium model is pretty good. But I use the 3Gb all day every day. I built a personal voice agent https://github.com/lawless-m/TheHand", "context": "", "use_case": "local speech recognition", "switched_from": "", "story": "Dell admits consumers don't care about AI PCs"}, {"id": 734901824, "item_id": 45931364, "url": "https://news.ycombinator.com/item?id=45931364", "by": "lxe", "date": "2025-11-14", "stance": "recommend", "quote": "I much prefer whisper after playing with both. Maybe I'm not configuring Parakeet correctly, But the transcription that comes out of Whisper is usually pretty much spot on.", "excerpt": "I've done something similar for Linux and Mac. I originally used Whisper and then switched to Parakeet. I much prefer whisper after playing with both. Maybe I'm not configuring Parakeet correctly, But the transcription that comes out of Whisper is usually pretty much spot on. It automatically removes all the \"ooms\" and all the \"ahs\" and it's just way more natural, in my opinion. I'm using Whisper.CPP with CUDA accele", "context": "built dictation tools for Linux and Mac", "use_case": "voice dictation", "switched_from": "Parakeet", "story": "Show HN: Chirp \u2013 Local Windows dictation with ParakeetV3 no executable required"}, {"id": 775384064, "item_id": 48461504, "url": "https://news.ycombinator.com/item?id=48461504", "by": "prometheus76", "date": "2026-06-09", "stance": "recommend", "quote": "I use a local instantiation of whisper to transcribe it when I get home.", "excerpt": "Relatedly, I have recently started walking and talking, and I just use lightweight gamer headphones that have a mic. I just record it with a generic voice recorder app, then I use a local instantiation of whisper to transcribe it when I get home. I throw the transcript and the recording into a database and then I can search the whole thing, or have AI look for interesting patterns in my thoughts.", "context": "", "use_case": "local voice transcription", "switched_from": "", "story": "Ask HN: What are tools you have made for yourself since the advent of AI"}]}, {"text": ", especially for dictation and multilingual transcription.", "cites": [{"id": 744973633, "item_id": 46560852, "url": "https://news.ycombinator.com/item?id=46560852", "by": "delaminator", "date": "2026-01-09", "stance": "recommend", "quote": "My favourite use is Whisper. I voice->text almost all of my typing now.", "excerpt": "I have a 3090 24Gb Twin Xenon 64Gb RAM sat on a machine in our server room. I do local AI with Qwen, Whisper and another I can't remember right now. These are all QWEN: We do AI Invoice OCR - PDF -> Image -> Excel. Works much better than other solutions because it has invoice context so it looks for particular data to extract and ignores others. Why local? I proved it worked, no need to send our data outside for proc", "context": "", "use_case": "voice dictation", "switched_from": "", "story": "Ask HN: Who's running local AI workstations in 2026?"}, {"id": 747873680, "item_id": 46742105, "url": "https://news.ycombinator.com/item?id=46742105", "by": "woudsma", "date": "2026-01-24", "stance": "recommend", "quote": "I\u2019m using whisper with superwhisper on my mac.", "excerpt": "I\u2019m using whisper with superwhisper on my mac. I\u2019ve assigned a key on my keyboard, when I press the key it starts listening and when I release it, the text gets copied to the current cursor location. It works pretty well.", "context": "", "use_case": "local dictation on Mac", "switched_from": "", "story": "Ask HN: What's the current best local/open speech-to-speech setup?"}, {"id": 797773936, "item_id": 49860871, "url": "https://news.ycombinator.com/item?id=49860871", "by": "gunalx", "date": "2026-09-26", "stance": "recommend", "quote": "Whisper is still pretty good over all. At least if you need multilingual support.", "excerpt": "Whisper is still pretty good over all. At least if you need multilingual support. Other mentions is parakeet for speed and maybe qwen3 asr.", "context": "", "use_case": "local speech recognition", "switched_from": "", "story": "Ask HN: What would be the best local ASR model"}, {"id": 802823696, "item_id": 50176481, "url": "https://news.ycombinator.com/item?id=50176481", "by": "throwaway219450", "date": "2026-10-21", "stance": "recommend", "quote": "I use Whisper to transcribe foreign language videos and then run them through a couple of models to translate into English. It works really well", "excerpt": "> Over the last few decades I\u2019ve read many technical docs that were obviously mechanically translated from some other language. This is much better. I don\u2019t think it\u2019s better. Any frontier LLM is capable of competent translation without significantly re-wording the text. I use Whisper to transcribe foreign language videos and then run them through a couple of models to translate into English. It works really well, an", "context": "", "use_case": "transcribing foreign language videos", "switched_from": "", "story": "IKEA \u2013 3D Guide"}]}, {"text": " The most common complaint is transcription mistakes and hallucinations", "cites": [{"id": 782335714, "item_id": 48895982, "url": "https://news.ycombinator.com/item?id=48895982", "by": "trencedamp", "date": "2026-07-13", "stance": "warn", "quote": "If I switch to whisper it makes lots of mistakes", "excerpt": "Came here to post this. I use handy on my own machine and it's perfect with parakeet. If I switch to whisper it makes lots of mistakes", "context": "", "use_case": "speech transcription", "switched_from": "", "story": "Apple's new SpeechAnalyzer API, benchmarked against Whisper and its predecessor"}, {"id": 782331761, "item_id": 48895735, "url": "https://news.ycombinator.com/item?id=48895735", "by": "summarity", "date": "2026-07-13", "stance": "warn", "quote": "Whisper was catastrophically bad.", "excerpt": "Vs Voxtral would be a better comparison. No other model, open or closed, has been able to hit such a low AER (Acronym Error Rate ;)) for my meeting transcripts. Seems to understand/infer all the technobabble I use at work. Never have to edit anything. Whisper was catastrophically bad.", "context": "", "use_case": "meeting transcription", "switched_from": "", "story": "Apple's new SpeechAnalyzer API, benchmarked against Whisper and its predecessor"}, {"id": 791650224, "item_id": 49478139, "url": "https://news.ycombinator.com/item?id=49478139", "by": "decide1000", "date": "2026-08-28", "stance": "neutral", "quote": "Whisper often makes mistakes which I now solve with an extra layer (qwen).", "excerpt": "I use Whisper in combination with Qwen3.8 27b to create Dutch subtitles for Russian videos. I follow Ukrainian channels about the war and I am now able to understand them. In my tests with other languages Voxtral was better then Whisper. Whisper often makes mistakes which I now solve with an extra layer (qwen). Sadly Mistral does not support Ukrainian. I hope this changes! I expect better from a European model", "context": "", "use_case": "Dutch subtitles for Russian videos", "switched_from": "", "story": "Gemini-3.5-Transcribe"}, {"id": 734830208, "item_id": 45926888, "url": "https://news.ycombinator.com/item?id=45926888", "by": "geerlingguy", "date": "2025-11-14", "stance": "recommend", "quote": "Whisper has radically improved CC on my video content.", "excerpt": "I like the conclusion; like for me, Whisper has radically improved CC on my video content. I used to spend a few hours translating my scripts into CCs, and tooling was poor. Now I run it through whisper in a couple minutes, give one quick pass to correct a few small hallucinations and misspellings, and I'm done. There are big wins in AI. But those don't pump the bubble once they're solved. And the thing that made Whi", "context": "creates video content", "use_case": "video captions transcription", "switched_from": "", "story": "AGI fantasy is a blocker to actual engineering"}]}, {"text": ", and some now prefer Parakeet for speed.", "cites": [{"id": 782324131, "item_id": 48895258, "url": "https://news.ycombinator.com/item?id=48895258", "by": "coder543", "date": "2026-07-13", "stance": "compare", "quote": "I would choose Parakeet over Whisper in almost all situations these days.", "excerpt": "At this point, I would not recommend ignoring Parakeet TDT 0.6b v2/v3 (english-only versus multilingual). Those models have been out for a year, give or take, and they are both accurate and fast. I would choose Parakeet over Whisper in almost all situations these days. Parakeet works great on my iPhone 15 Pro Max, so if an app is going to ship a dedicated model, I strongly recommend investigating Parakeet. On the mor", "context": "", "use_case": "speech transcription", "switched_from": "", "story": "Apple's new SpeechAnalyzer API, benchmarked against Whisper and its predecessor"}, {"id": 750319986, "item_id": 46894999, "url": "https://news.ycombinator.com/item?id=46894999", "by": "d4rkp4ttern", "date": "2026-02-04", "stance": "compare", "quote": "accuracy slightly less than slower Whisper models", "excerpt": "I\u2019m curious about this too. On my M1 Max MacBook I use the Handy app on macOS with Parakeet V3 and I get near instant transcription, accuracy slightly less than slower Whisper models, but that drop is immaterial when talking to CLI coding agents, which is where I find the most use for this. https://github.com/cjpais/Handy", "context": "", "use_case": "transcription", "switched_from": "", "story": "Voxtral Transcribe 2"}, {"id": 797773936, "item_id": 49860871, "url": "https://news.ycombinator.com/item?id=49860871", "by": "gunalx", "date": "2026-09-26", "stance": "recommend", "quote": "Whisper is still pretty good over all. At least if you need multilingual support.", "excerpt": "Whisper is still pretty good over all. At least if you need multilingual support. Other mentions is parakeet for speed and maybe qwen3 asr.", "context": "", "use_case": "local speech recognition", "switched_from": "", "story": "Ask HN: What would be the best local ASR model"}]}], "praise": [{"label": "Runs locally", "summary": "Commenters run it on their own machines, phones and home servers without sending data to the cloud.", "quotes": [{"id": 782890065, "item_id": 48930629, "url": "https://news.ycombinator.com/item?id=48930629", "by": "wolvoleo", "date": "2026-07-15", "stance": "recommend", "quote": "a locally hosted whisper instance backed by Qwen 3.6", "excerpt": "I kinda have that though. I use a home assistant voice preview speaker, connected to a locally hosted whisper instance backed by Qwen 3.6. All local and fairly energy-reasonable :) Just saying it's not impossible to have your cake and eat it.", "context": "", "use_case": "local speech recognition", "switched_from": "", "story": "Guerrilla London Bus Ads Mock Kylie Jenner's Meta Glasses Campaign"}, {"id": 775384064, "item_id": 48461504, "url": "https://news.ycombinator.com/item?id=48461504", "by": "prometheus76", "date": "2026-06-09", "stance": "recommend", "quote": "I use a local instantiation of whisper to transcribe it when I get home.", "excerpt": "Relatedly, I have recently started walking and talking, and I just use lightweight gamer headphones that have a mic. I just record it with a generic voice recorder app, then I use a local instantiation of whisper to transcribe it when I get home. I throw the transcript and the recording into a database and then I can search the whole thing, or have AI look for interesting patterns in my thoughts.", "context": "", "use_case": "local voice transcription", "switched_from": "", "story": "Ask HN: What are tools you have made for yourself since the advent of AI"}, {"id": 755817061, "item_id": 47238566, "url": "https://news.ycombinator.com/item?id=47238566", "by": "ali_chherawalla", "date": "2026-03-03", "stance": "neutral", "quote": "Voice is on-device Whisper.", "excerpt": "ofcourse ofcourse, I've documented everything here: https://github.com/alichherawalla/off-grid-mobile-ai/blob/main/docs/ARCHITECTURE.md llama.cpp compiled as a native Android library via the NDK, linked into React Native through a custom JSI bridge. GGUF models loaded straight into memory. On Snapdragon devices we use QNN (Qualcomm Neural Network) for hardware acceleration. OpenCL GPU fallback on everything else. CPU", "context": "built an on-device Android AI app", "use_case": "on-device voice", "switched_from": "", "story": "Show HN: Qwen 3.5 running on a $300 Android phone \u2013 on-device, open source"}]}, {"label": "Accuracy", "summary": "Several commenters say its transcriptions are accurate, often better than built-in OS speech-to-text.", "quotes": [{"id": 734901824, "item_id": 45931364, "url": "https://news.ycombinator.com/item?id=45931364", "by": "lxe", "date": "2025-11-14", "stance": "recommend", "quote": "I much prefer whisper after playing with both. Maybe I'm not configuring Parakeet correctly, But the transcription that comes out of Whisper is usually pretty much spot on.", "excerpt": "I've done something similar for Linux and Mac. I originally used Whisper and then switched to Parakeet. I much prefer whisper after playing with both. Maybe I'm not configuring Parakeet correctly, But the transcription that comes out of Whisper is usually pretty much spot on. It automatically removes all the \"ooms\" and all the \"ahs\" and it's just way more natural, in my opinion. I'm using Whisper.CPP with CUDA accele", "context": "built dictation tools for Linux and Mac", "use_case": "voice dictation", "switched_from": "Parakeet", "story": "Show HN: Chirp \u2013 Local Windows dictation with ParakeetV3 no executable required"}, {"id": 767643826, "item_id": 47977739, "url": "https://news.ycombinator.com/item?id=47977739", "by": "nerdsniper", "date": "2026-05-01", "stance": "recommend", "quote": "Whisper models in general seem to do a lot better than most built-into-OS/app models.", "excerpt": "On just the transcription side: WhisprFlow produces much better speech-to-text for long text messaging-by-voice (dictation / transcription) than apple's speech-to-text does. Whisper models in general seem to do a lot better than most built-into-OS/app models. Which is interesting, because there's nothing stopping them from just using Whisper models. I love MacWhisper personally. Also, Gumroad is a fantastic app distr", "context": "", "use_case": "speech-to-text", "switched_from": "", "story": "Apple accidentally left Claude.md files Apple Support app"}, {"id": 751940768, "item_id": 46996298, "url": "https://news.ycombinator.com/item?id=46996298", "by": "srinifromsalem", "date": "2026-02-12", "stance": "recommend", "quote": "Whisper remains the gold standard - you can run it locally with good accuracy across languages.", "excerpt": "For local speech-to-text, Whisper remains the gold standard - you can run it locally with good accuracy across languages. For speech-to-speech, you'd typically chain Whisper with a local TTS model like Coqui TTS or use something like Tortoise TTS for higher quality but slower processing. The key is balancing accuracy, speed, and resource usage based on your specific use case. If you're doing content creation workflow", "context": "", "use_case": "local speech-to-text", "switched_from": "", "story": "Ask HN: What's the current best local/open speech-to-speech setup?"}]}, {"label": "Dictation", "summary": "People use it to dictate much of their typing, often through hotkey apps on the Mac.", "quotes": [{"id": 744973633, "item_id": 46560852, "url": "https://news.ycombinator.com/item?id=46560852", "by": "delaminator", "date": "2026-01-09", "stance": "recommend", "quote": "My favourite use is Whisper. I voice->text almost all of my typing now.", "excerpt": "I have a 3090 24Gb Twin Xenon 64Gb RAM sat on a machine in our server room. I do local AI with Qwen, Whisper and another I can't remember right now. These are all QWEN: We do AI Invoice OCR - PDF -> Image -> Excel. Works much better than other solutions because it has invoice context so it looks for particular data to extract and ignores others. Why local? I proved it worked, no need to send our data outside for proc", "context": "", "use_case": "voice dictation", "switched_from": "", "story": "Ask HN: Who's running local AI workstations in 2026?"}, {"id": 747873680, "item_id": 46742105, "url": "https://news.ycombinator.com/item?id=46742105", "by": "woudsma", "date": "2026-01-24", "stance": "recommend", "quote": "I\u2019m using whisper with superwhisper on my mac.", "excerpt": "I\u2019m using whisper with superwhisper on my mac. I\u2019ve assigned a key on my keyboard, when I press the key it starts listening and when I release it, the text gets copied to the current cursor location. It works pretty well.", "context": "", "use_case": "local dictation on Mac", "switched_from": "", "story": "Ask HN: What's the current best local/open speech-to-speech setup?"}, {"id": 735736304, "item_id": 45983519, "url": "https://news.ycombinator.com/item?id=45983519", "by": "mesadb", "date": "2025-11-19", "stance": "recommend", "quote": "Current version sticks to FFmpeg + Whisper CLI locally, optimized for short chunks so you usually see text within about 1\u20132 seconds.", "excerpt": "Some implementation details, since getting this to work well was not trivial. My goal was \u201cpress hotkey, start talking, see text within ~1\u20132 seconds\u201d on an M2 MacBook Pro, and support multiple languages. First attempts (cloud) \u2013 I tried Hugging Face real-time transcription. It worked but latency was all over the place and costs would not scale. \u2013 I tried OpenAI real-time transcription. Latency was better, but when th", "context": "built a macOS dictation app", "use_case": "local speech transcription", "switched_from": "OpenAI real-time transcription", "story": "Show HN: Slave \u2013 local dictation and TTS for macOS (3k words free)"}]}, {"label": "Multilingual", "summary": "Valued for handling many languages, including transcribing foreign videos for translation.", "quotes": [{"id": 797773936, "item_id": 49860871, "url": "https://news.ycombinator.com/item?id=49860871", "by": "gunalx", "date": "2026-09-26", "stance": "recommend", "quote": "Whisper is still pretty good over all. At least if you need multilingual support.", "excerpt": "Whisper is still pretty good over all. At least if you need multilingual support. Other mentions is parakeet for speed and maybe qwen3 asr.", "context": "", "use_case": "local speech recognition", "switched_from": "", "story": "Ask HN: What would be the best local ASR model"}, {"id": 802823696, "item_id": 50176481, "url": "https://news.ycombinator.com/item?id=50176481", "by": "throwaway219450", "date": "2026-10-21", "stance": "recommend", "quote": "I use Whisper to transcribe foreign language videos and then run them through a couple of models to translate into English. It works really well", "excerpt": "> Over the last few decades I\u2019ve read many technical docs that were obviously mechanically translated from some other language. This is much better. I don\u2019t think it\u2019s better. Any frontier LLM is capable of competent translation without significantly re-wording the text. I use Whisper to transcribe foreign language videos and then run them through a couple of models to translate into English. It works really well, an", "context": "", "use_case": "transcribing foreign language videos", "switched_from": "", "story": "IKEA \u2013 3D Guide"}, {"id": 751940768, "item_id": 46996298, "url": "https://news.ycombinator.com/item?id=46996298", "by": "srinifromsalem", "date": "2026-02-12", "stance": "recommend", "quote": "Whisper remains the gold standard - you can run it locally with good accuracy across languages.", "excerpt": "For local speech-to-text, Whisper remains the gold standard - you can run it locally with good accuracy across languages. For speech-to-speech, you'd typically chain Whisper with a local TTS model like Coqui TTS or use something like Tortoise TTS for higher quality but slower processing. The key is balancing accuracy, speed, and resource usage based on your specific use case. If you're doing content creation workflow", "context": "", "use_case": "local speech-to-text", "switched_from": "", "story": "Ask HN: What's the current best local/open speech-to-speech setup?"}]}, {"label": "Video captions", "summary": "Creators use it to generate captions and subtitles in minutes instead of hours.", "quotes": [{"id": 734830208, "item_id": 45926888, "url": "https://news.ycombinator.com/item?id=45926888", "by": "geerlingguy", "date": "2025-11-14", "stance": "recommend", "quote": "Whisper has radically improved CC on my video content.", "excerpt": "I like the conclusion; like for me, Whisper has radically improved CC on my video content. I used to spend a few hours translating my scripts into CCs, and tooling was poor. Now I run it through whisper in a couple minutes, give one quick pass to correct a few small hallucinations and misspellings, and I'm done. There are big wins in AI. But those don't pump the bubble once they're solved. And the thing that made Whi", "context": "creates video content", "use_case": "video captions transcription", "switched_from": "", "story": "AGI fantasy is a blocker to actual engineering"}, {"id": 791650224, "item_id": 49478139, "url": "https://news.ycombinator.com/item?id=49478139", "by": "decide1000", "date": "2026-08-28", "stance": "neutral", "quote": "Whisper often makes mistakes which I now solve with an extra layer (qwen).", "excerpt": "I use Whisper in combination with Qwen3.8 27b to create Dutch subtitles for Russian videos. I follow Ukrainian channels about the war and I am now able to understand them. In my tests with other languages Voxtral was better then Whisper. Whisper often makes mistakes which I now solve with an extra layer (qwen). Sadly Mistral does not support Ukrainian. I hope this changes! I expect better from a European model", "context": "", "use_case": "Dutch subtitles for Russian videos", "switched_from": "", "story": "Gemini-3.5-Transcribe"}]}, {"label": "Building block", "summary": "Seen as a small, specialized model that is easy to stack into larger pipelines.", "quotes": [{"id": 773500016, "item_id": 48343751, "url": "https://news.ycombinator.com/item?id=48343751", "by": "sureglymop", "date": "2026-05-31", "stance": "recommend", "quote": "smaller specialised things like whisper, CLIP, SAM, etc.", "excerpt": "What I'm actually excited about in AI are smaller specialised things like whisper, CLIP, SAM, etc. Just stacking them together as little building blocks has been amazing. Great app to hopefully let more people realize that it's not all just huge LLMs that have to run somewhere else!", "context": "", "use_case": "speech recognition building block", "switched_from": "", "story": "Ask HN: What Is the State of App Development in 2026?"}, {"id": 747854353, "item_id": 46740897, "url": "https://news.ycombinator.com/item?id=46740897", "by": "benatkin", "date": "2026-01-23", "stance": "recommend", "quote": "it would make more sense to go straight to Whisper.", "excerpt": "To save a click, it's just a fancy front end for Whisper plus a weaker CPU-only model. It has a demo video that seems impressive, but the speech is careful to sound casual while having no meaningful flaws that would cause it to mess up. If you want to make a speech to speech tool, which is what this post asks about, it would make more sense to go straight to Whisper.", "context": "", "use_case": "speech-to-speech tooling", "switched_from": "", "story": "Ask HN: What's the current best local/open speech-to-speech setup?"}, {"id": 783751105, "item_id": 48984444, "url": "https://news.ycombinator.com/item?id=48984444", "by": "TehCorwiz", "date": "2026-07-20", "stance": "neutral", "quote": "audiobook -> text (via whisper)", "excerpt": "I used Claude to build a complete data extraction pipeline for a popular current best seller book series: audiobook -> text (via whisper) -> local LLM (qwen) -> database. Not once did it seem to acknowledge or care about copyright. It even used knowledge it already had about the books to exclude certain ones before beginning since the character I was interested in did not appear in those. It definitely had context of", "context": "", "use_case": "audiobook transcription", "switched_from": "", "story": "American AI is locked down and proprietary. It's losing"}]}], "caveats": [{"label": "Transcription errors", "summary": "Some report frequent mistakes or hallucinations that need correcting or a second model pass.", "quotes": [{"id": 782335714, "item_id": 48895982, "url": "https://news.ycombinator.com/item?id=48895982", "by": "trencedamp", "date": "2026-07-13", "stance": "warn", "quote": "If I switch to whisper it makes lots of mistakes", "excerpt": "Came here to post this. I use handy on my own machine and it's perfect with parakeet. If I switch to whisper it makes lots of mistakes", "context": "", "use_case": "speech transcription", "switched_from": "", "story": "Apple's new SpeechAnalyzer API, benchmarked against Whisper and its predecessor"}, {"id": 782331761, "item_id": 48895735, "url": "https://news.ycombinator.com/item?id=48895735", "by": "summarity", "date": "2026-07-13", "stance": "warn", "quote": "Whisper was catastrophically bad.", "excerpt": "Vs Voxtral would be a better comparison. No other model, open or closed, has been able to hit such a low AER (Acronym Error Rate ;)) for my meeting transcripts. Seems to understand/infer all the technobabble I use at work. Never have to edit anything. Whisper was catastrophically bad.", "context": "", "use_case": "meeting transcription", "switched_from": "", "story": "Apple's new SpeechAnalyzer API, benchmarked against Whisper and its predecessor"}, {"id": 791650224, "item_id": 49478139, "url": "https://news.ycombinator.com/item?id=49478139", "by": "decide1000", "date": "2026-08-28", "stance": "neutral", "quote": "Whisper often makes mistakes which I now solve with an extra layer (qwen).", "excerpt": "I use Whisper in combination with Qwen3.8 27b to create Dutch subtitles for Russian videos. I follow Ukrainian channels about the war and I am now able to understand them. In my tests with other languages Voxtral was better then Whisper. Whisper often makes mistakes which I now solve with an extra layer (qwen). Sadly Mistral does not support Ukrainian. I hope this changes! I expect better from a European model", "context": "", "use_case": "Dutch subtitles for Russian videos", "switched_from": "", "story": "Gemini-3.5-Transcribe"}]}, {"label": "Parakeet faster", "summary": "Several commenters prefer Parakeet for speed, trading a little accuracy.", "quotes": [{"id": 782324131, "item_id": 48895258, "url": "https://news.ycombinator.com/item?id=48895258", "by": "coder543", "date": "2026-07-13", "stance": "compare", "quote": "I would choose Parakeet over Whisper in almost all situations these days.", "excerpt": "At this point, I would not recommend ignoring Parakeet TDT 0.6b v2/v3 (english-only versus multilingual). Those models have been out for a year, give or take, and they are both accurate and fast. I would choose Parakeet over Whisper in almost all situations these days. Parakeet works great on my iPhone 15 Pro Max, so if an app is going to ship a dedicated model, I strongly recommend investigating Parakeet. On the mor", "context": "", "use_case": "speech transcription", "switched_from": "", "story": "Apple's new SpeechAnalyzer API, benchmarked against Whisper and its predecessor"}, {"id": 750319986, "item_id": 46894999, "url": "https://news.ycombinator.com/item?id=46894999", "by": "d4rkp4ttern", "date": "2026-02-04", "stance": "compare", "quote": "accuracy slightly less than slower Whisper models", "excerpt": "I\u2019m curious about this too. On my M1 Max MacBook I use the Handy app on macOS with Parakeet V3 and I get near instant transcription, accuracy slightly less than slower Whisper models, but that drop is immaterial when talking to CLI coding agents, which is where I find the most use for this. https://github.com/cjpais/Handy", "context": "", "use_case": "transcription", "switched_from": "", "story": "Voxtral Transcribe 2"}, {"id": 797773936, "item_id": 49860871, "url": "https://news.ycombinator.com/item?id=49860871", "by": "gunalx", "date": "2026-09-26", "stance": "recommend", "quote": "Whisper is still pretty good over all. At least if you need multilingual support.", "excerpt": "Whisper is still pretty good over all. At least if you need multilingual support. Other mentions is parakeet for speed and maybe qwen3 asr.", "context": "", "use_case": "local speech recognition", "switched_from": "", "story": "Ask HN: What would be the best local ASR model"}]}, {"label": "Multiple speakers", "summary": "Works well for one speaker, but more speakers or complex repetitions produce repeated or poor output.", "quotes": [{"id": 747879313, "item_id": 46742457, "url": "https://news.ycombinator.com/item?id=46742457", "by": "dvfjsdhgfv", "date": "2026-01-24", "stance": "recommend", "quote": "Whisper is really good for 1 person speaking.", "excerpt": "I spent a few days on similar scenario without much success (scenario where one person speaks and then their speech is translated, and I want juts the original or both). An API call to GPT4o works quite well (it basically handles both transcription and diarization), but I wanted a local model. Whisper is really good for 1 person speaking. With more people you get repetitions. Qwen and other open multimodal models giv", "context": "", "use_case": "speech-to-text", "switched_from": "", "story": "Ask HN: What's the current best local/open speech-to-speech setup?"}, {"id": 733945317, "item_id": 45871582, "url": "https://news.ycombinator.com/item?id=45871582", "by": "wasi0013", "date": "2025-11-09", "stance": "warn", "quote": "the audio files have some complex repetitions; the output was miserable.", "excerpt": "Experimenting with various AI models via GitHub Co-Pilot on an extremely niche project to see how far these models have progressed. Used like ~60% of the premium quota to develop the following projects: website: https://murajah.pages.dev/ Play Store: https://play.google.com/store/apps/details?id=com.murajah.webview Manual audio splitting tool for the above project: audio-splitter-6b3.pages.dev/ I've always been skept", "context": "", "use_case": "automatic audio splitting", "switched_from": "", "story": "Ask HN: What Are You Working On? (Nov 2025)"}]}, {"label": "Heavy to run", "summary": "One user found it too heavy to run locally at good quality.", "quotes": [{"id": 782441249, "item_id": 48902578, "url": "https://news.ycombinator.com/item?id=48902578", "by": "Freddy153", "date": "2026-07-13", "stance": "warn", "quote": "I find Whisper too heavy to run locally with good quality", "excerpt": "I have implemented this into Home Assustant some months (https://github.com/FI-153/wyoming-apple-stt). I find the comparison with Whisper particularly helpful since it is the base model used in Home Assistant to transcribe voice in the Assist pipeline. I find Whisper too heavy to run locally with good quality, SpeechAnalyzer instead comes bundled in macOS and is always available.", "context": "Home Assistant user", "use_case": "local transcription", "switched_from": "", "story": "Apple's new SpeechAnalyzer API, benchmarked against Whisper and its predecessor"}]}], "who": [{"by": "ali_chherawalla", "context": "built an on-device Android AI app", "url": "https://news.ycombinator.com/item?id=47238566", "date": "2026-03-03"}, {"by": "cbrews", "context": "audio background", "url": "https://news.ycombinator.com/item?id=46739544", "date": "2026-01-23"}, {"by": "mesadb", "context": "built a macOS dictation app", "url": "https://news.ycombinator.com/item?id=45983519", "date": "2025-11-19"}, {"by": "lxe", "context": "built dictation tools for Linux and Mac", "url": "https://news.ycombinator.com/item?id=45931364", "date": "2025-11-14"}, {"by": "geerlingguy", "context": "creates video content", "url": "https://news.ycombinator.com/item?id=45926888", "date": "2025-11-14"}], "tags": ["self-hostable", "for dictation", "for Mac users", "for video subtitles", "for multilingual audio"], "use_cases": ["speech transcription", "speech-to-text", "speech recognition", "local speech recognition", "video transcription", "voice dictation"], "compared": [{"slug": "parakeet", "name": "Parakeet", "co": 10, "switched_from": 1, "switched_to": 1}, {"slug": "handy", "name": "Handy", "co": 7, "switched_from": 0, "switched_to": 0}, {"slug": "claude-code", "name": "Claude Code", "co": 6, "switched_from": 0, "switched_to": 0}, {"slug": "voxtral", "name": "Voxtral", "co": 2, "switched_from": 0, "switched_to": 1}, {"slug": "qwen", "name": "Qwen", "co": 4, "switched_from": 0, "switched_to": 0}, {"slug": "claude-sonnet", "name": "Claude Sonnet", "co": 3, "switched_from": 0, "switched_to": 0}, {"slug": "ollama", "name": "Ollama", "co": 3, "switched_from": 0, "switched_to": 0}], "makers": [{"id": 792075280, "item_id": 49504705, "url": "https://news.ycombinator.com/item?id=49504705", "by": "smoores", "date": "2026-08-31", "stance": "neutral", "quote": "We use Whisper to transcribe the audio to text, and then use error-align[1] to align on the text.", "excerpt": "Oh, cool! Yeah that seems like a good application. The current Storyteller alignment algorithm actually does do just that! We use Whisper to transcribe the audio to text, and then use error-align[1] to align on the text. There are a few disadvantages to this approach: 1. Whisper only supports ~25 languages, and only about 10 of those very well. We want to support more languages, and Massively Multilingual Speech supp", "context": "", "use_case": "audio transcription for alignment", "switched_from": "", "story": "Automating Immersive Reading"}], "threads": [{"url": "https://news.ycombinator.com/item?id=46731068", "title": "Ask HN: What's the current best local/open speech-to-speech setup?", "mentions": 7}, {"url": "https://news.ycombinator.com/item?id=48894752", "title": "Apple's new SpeechAnalyzer API, benchmarked against Whisper and its predecessor", "mentions": 4}, {"url": "https://news.ycombinator.com/item?id=50168096", "title": "IKEA \u2013 3D Guide", "mentions": 1}, {"url": "https://news.ycombinator.com/item?id=50018918", "title": "Ask HN: How do you turn videos into searchable text without uploading them?", "mentions": 1}, {"url": "https://news.ycombinator.com/item?id=50008427", "title": "Whistle: Speech to Text in 16.9 MB", "mentions": 1}, {"url": "https://news.ycombinator.com/item?id=49922568", "title": "Ask HN: Who wants to be hired? (October 2026)", "mentions": 1}], "sentiment_by_month": {"2025-11": {"warn": 1, "recommend": 3}, "2025-12": {}, "2026-01": {"neutral": 3, "recommend": 4, "compare": 1}, "2026-02": {"compare": 1, "neutral": 1, "recommend": 1}, "2026-03": {"neutral": 2}, "2026-04": {}, "2026-05": {"recommend": 2, "neutral": 2}, "2026-06": {"neutral": 1, "recommend": 1}, "2026-07": {"neutral": 4, "compare": 1, "warn": 3, "recommend": 1}, "2026-08": {"neutral": 2}, "2026-09": {"neutral": 1, "recommend": 2}, "2026-10": {"neutral": 3, "recommend": 1}}, "quotes": [{"id": 802823696, "item_id": 50176481, "url": "https://news.ycombinator.com/item?id=50176481", "by": "throwaway219450", "date": "2026-10-21", "stance": "recommend", "quote": "I use Whisper to transcribe foreign language videos and then run them through a couple of models to translate into English. It works really well", "excerpt": "> Over the last few decades I\u2019ve read many technical docs that were obviously mechanically translated from some other language. This is much better. I don\u2019t think it\u2019s better. Any frontier LLM is capable of competent translation without significantly re-wording the text. I use Whisper to transcribe foreign language videos and then run them through a couple of models to translate into English. It works really well, an", "context": "", "use_case": "transcribing foreign language videos", "switched_from": "", "story": "IKEA \u2013 3D Guide"}, {"id": 800302944, "item_id": 50018934, "url": "https://news.ycombinator.com/item?id=50018934", "by": "apresmoi", "date": "2026-10-09", "stance": "neutral", "quote": "load whisper there and use the cloud GPUs to process them", "excerpt": "I have some Google Collab notebooks where I download/upload the videos, load whisper there and use the cloud GPUs to process them! Here is an example: https://github.com/apresmoi/jianglens/blob/main/ops/notebooks/colab/Whisper_Transcription.ipynb", "context": "", "use_case": "video transcription", "switched_from": "", "story": "Ask HN: How do you turn videos into searchable text without uploading them?"}, {"id": 800177296, "item_id": 50011081, "url": "https://news.ycombinator.com/item?id=50011081", "by": "raddan", "date": "2026-10-08", "stance": "neutral", "quote": "When I use Whisper in the terminal, I can see it going back and correcting itself.", "excerpt": "What do streaming implementations do when a bigger context reveals a different interpretation/parse? When I use Whisper in the terminal, I can see it going back and correcting itself. Are corrections off the table for a true streaming transcription?", "context": "", "use_case": "speech transcription", "switched_from": "", "story": "Whistle: Speech to Text in 16.9 MB"}, {"id": 798894226, "item_id": 49930889, "url": "https://news.ycombinator.com/item?id=49930889", "by": "Francisco-C", "date": "2026-10-02", "stance": "neutral", "quote": "It used Whisper STT, Azure Neural TTS containers, SSML/prosody control", "excerpt": "Location: Coimbra, Portugal Remote: Yes Willing to relocate: No Technologies: Python, TypeScript, C#, .NET 8, FastAPI, Node.js, React, Next.js, PostgreSQL, Redis, MongoDB, Azure, Docker, OpenAI APIs, Claude Code, Codex, RAG, agentic workflows R\u00e9sum\u00e9/CV: https://github.com/inclyner/cv Email: CarvalhoFrancisco@proton.me Early-career software engineer working remotely for a Swiss startup. I have around 9 months of full-", "context": "", "use_case": "speech-to-text in dubbing pipeline", "switched_from": "", "story": "Ask HN: Who wants to be hired? (October 2026)"}, {"id": 797773936, "item_id": 49860871, "url": "https://news.ycombinator.com/item?id=49860871", "by": "gunalx", "date": "2026-09-26", "stance": "recommend", "quote": "Whisper is still pretty good over all. At least if you need multilingual support.", "excerpt": "Whisper is still pretty good over all. At least if you need multilingual support. Other mentions is parakeet for speed and maybe qwen3 asr.", "context": "", "use_case": "local speech recognition", "switched_from": "", "story": "Ask HN: What would be the best local ASR model"}, {"id": 795951681, "item_id": 49746980, "url": "https://news.ycombinator.com/item?id=49746980", "by": "arnav714412", "date": "2026-09-17", "stance": "recommend", "quote": "Kimmi with whisper.", "excerpt": "Kimmi with whisper. Ngl running it on my old laptop is the only thing keeping my thing from breaking apart", "context": "", "use_case": "speech transcription", "switched_from": "", "story": "Ask HN: What self-hosted apps do you use daily?"}, {"id": 793600994, "item_id": 49600062, "url": "https://news.ycombinator.com/item?id=49600062", "by": "Francisco-C", "date": "2026-09-07", "stance": "neutral", "quote": "It used Whisper STT, Azure Neural TTS containers", "excerpt": "Location: Coimbra, Portugal Remote: Yes Willing to relocate: No Technologies: Python, TypeScript, C#, .NET 8, FastAPI, Node.js, React, Next.js, PostgreSQL, Redis, MongoDB, Azure, Docker, OpenAI APIs, Claude Code, Codex, RAG, agentic workflows R\u00e9sum\u00e9/CV: https://github.com/inclyner/cv Email: CarvalhoFrancisco@proton.me Early-career software engineer working remotely for a Swiss startup. I have around 9 months of full-", "context": "", "use_case": "video dubbing speech-to-text", "switched_from": "", "story": "Ask HN: Who wants to be hired? (September 2026)"}, {"id": 791650224, "item_id": 49478139, "url": "https://news.ycombinator.com/item?id=49478139", "by": "decide1000", "date": "2026-08-28", "stance": "neutral", "quote": "Whisper often makes mistakes which I now solve with an extra layer (qwen).", "excerpt": "I use Whisper in combination with Qwen3.8 27b to create Dutch subtitles for Russian videos. I follow Ukrainian channels about the war and I am now able to understand them. In my tests with other languages Voxtral was better then Whisper. Whisper often makes mistakes which I now solve with an extra layer (qwen). Sadly Mistral does not support Ukrainian. I hope this changes! I expect better from a European model", "context": "", "use_case": "Dutch subtitles for Russian videos", "switched_from": "", "story": "Gemini-3.5-Transcribe"}, {"id": 790955506, "item_id": 49434719, "url": "https://news.ycombinator.com/item?id=49434719", "by": "jlokier", "date": "2026-08-25", "stance": "neutral", "quote": "I used to use Ollama and Whisper", "excerpt": "You can't run Linux directly on these. Asahi Linux supports up to M2 only. Linux runs very well in a VM on macOS. There are many good options for this, some free and open source (QEMU, UTM, Lima, Colima), some proprietary (VMware Fusion, Parallels). But Linux in a VM doesn't get access to the real GPU, so model performance is limited. Those running on the CPU perform well, and those needing the GPU don't. However, ma", "context": "", "use_case": "speech recognition", "switched_from": "", "story": "Apple introduces M6 and M5 Ultra"}, {"id": 784425749, "item_id": 49026609, "url": "https://news.ycombinator.com/item?id=49026609", "by": "AmazingTurtle", "date": "2026-07-23", "stance": "neutral", "quote": "It\u2019s built with Tauri, a Rust backend, React/TypeScript, SQLite, GStreamer, Whisper, OCR, and VLM processing.", "excerpt": "Funny timing. I've been building something similar in my spare time called Daydream. There\u2019s a lot of overlap: local screen/audio capture, OCR and transcription, window and activity context, SQLite, and a searchable memory of what happened. The main difference is the product direction. Screenpipe seems focused on continuously giving agents context through APIs, MCP, and skills. Daydream is more narrowly built around ", "context": "", "use_case": "audio transcription", "switched_from": "", "story": "Launch HN: Screenpipe (YC S26) \u2013 Power your agents by your 24/7 screen recording"}, {"id": 783751105, "item_id": 48984444, "url": "https://news.ycombinator.com/item?id=48984444", "by": "TehCorwiz", "date": "2026-07-20", "stance": "neutral", "quote": "audiobook -> text (via whisper)", "excerpt": "I used Claude to build a complete data extraction pipeline for a popular current best seller book series: audiobook -> text (via whisper) -> local LLM (qwen) -> database. Not once did it seem to acknowledge or care about copyright. It even used knowledge it already had about the books to exclude certain ones before beginning since the character I was interested in did not appear in those. It definitely had context of", "context": "", "use_case": "audiobook transcription", "switched_from": "", "story": "American AI is locked down and proprietary. It's losing"}, {"id": 783405251, "item_id": 48962828, "url": "https://news.ycombinator.com/item?id=48962828", "by": "dandaka", "date": "2026-07-18", "stance": "neutral", "quote": "I use Gemma, Whisper and Qwen", "excerpt": "Few cases I have found very useful myself 1/ Using GUI software. My agents are using headful Google Chrome and Figma. It helps a lot to have separate environment, which is not interfering my main machine. 2/ Running long processes (1h+), so I can leave main machine closed. 3/ Running intensive processes. I use Gemma, Whisper and Qwen, which could burn main machine CPU and resources.", "context": "", "use_case": "local intensive processes", "switched_from": "", "story": "Setting up your spare Mac for Claude Code to control, a step-by-step guide"}, {"id": 782890065, "item_id": 48930629, "url": "https://news.ycombinator.com/item?id=48930629", "by": "wolvoleo", "date": "2026-07-15", "stance": "recommend", "quote": "a locally hosted whisper instance backed by Qwen 3.6", "excerpt": "I kinda have that though. I use a home assistant voice preview speaker, connected to a locally hosted whisper instance backed by Qwen 3.6. All local and fairly energy-reasonable :) Just saying it's not impossible to have your cake and eat it.", "context": "", "use_case": "local speech recognition", "switched_from": "", "story": "Guerrilla London Bus Ads Mock Kylie Jenner's Meta Glasses Campaign"}, {"id": 782441249, "item_id": 48902578, "url": "https://news.ycombinator.com/item?id=48902578", "by": "Freddy153", "date": "2026-07-13", "stance": "warn", "quote": "I find Whisper too heavy to run locally with good quality", "excerpt": "I have implemented this into Home Assustant some months (https://github.com/FI-153/wyoming-apple-stt). I find the comparison with Whisper particularly helpful since it is the base model used in Home Assistant to transcribe voice in the Assist pipeline. I find Whisper too heavy to run locally with good quality, SpeechAnalyzer instead comes bundled in macOS and is always available.", "context": "Home Assistant user", "use_case": "local transcription", "switched_from": "", "story": "Apple's new SpeechAnalyzer API, benchmarked against Whisper and its predecessor"}, {"id": 782335714, "item_id": 48895982, "url": "https://news.ycombinator.com/item?id=48895982", "by": "trencedamp", "date": "2026-07-13", "stance": "warn", "quote": "If I switch to whisper it makes lots of mistakes", "excerpt": "Came here to post this. I use handy on my own machine and it's perfect with parakeet. If I switch to whisper it makes lots of mistakes", "context": "", "use_case": "speech transcription", "switched_from": "", "story": "Apple's new SpeechAnalyzer API, benchmarked against Whisper and its predecessor"}, {"id": 782331761, "item_id": 48895735, "url": "https://news.ycombinator.com/item?id=48895735", "by": "summarity", "date": "2026-07-13", "stance": "warn", "quote": "Whisper was catastrophically bad.", "excerpt": "Vs Voxtral would be a better comparison. No other model, open or closed, has been able to hit such a low AER (Acronym Error Rate ;)) for my meeting transcripts. Seems to understand/infer all the technobabble I use at work. Never have to edit anything. Whisper was catastrophically bad.", "context": "", "use_case": "meeting transcription", "switched_from": "", "story": "Apple's new SpeechAnalyzer API, benchmarked against Whisper and its predecessor"}, {"id": 782324131, "item_id": 48895258, "url": "https://news.ycombinator.com/item?id=48895258", "by": "coder543", "date": "2026-07-13", "stance": "compare", "quote": "I would choose Parakeet over Whisper in almost all situations these days.", "excerpt": "At this point, I would not recommend ignoring Parakeet TDT 0.6b v2/v3 (english-only versus multilingual). Those models have been out for a year, give or take, and they are both accurate and fast. I would choose Parakeet over Whisper in almost all situations these days. Parakeet works great on my iPhone 15 Pro Max, so if an app is going to ship a dedicated model, I strongly recommend investigating Parakeet. On the mor", "context": "", "use_case": "speech transcription", "switched_from": "", "story": "Apple's new SpeechAnalyzer API, benchmarked against Whisper and its predecessor"}, {"id": 780450592, "item_id": 48778162, "url": "https://news.ycombinator.com/item?id=48778162", "by": "bobkb", "date": "2026-07-03", "stance": "neutral", "quote": "The whisper setup is something similar to what we have been using.", "excerpt": "Very useful. The whisper setup is something similar to what we have been using. The LLM setup though is outstanding.", "context": "", "use_case": "speech transcription", "switched_from": "", "story": "Jamesob's guide to running SOTA LLMs locally"}, {"id": 775384064, "item_id": 48461504, "url": "https://news.ycombinator.com/item?id=48461504", "by": "prometheus76", "date": "2026-06-09", "stance": "recommend", "quote": "I use a local instantiation of whisper to transcribe it when I get home.", "excerpt": "Relatedly, I have recently started walking and talking, and I just use lightweight gamer headphones that have a mic. I just record it with a generic voice recorder app, then I use a local instantiation of whisper to transcribe it when I get home. I throw the transcript and the recording into a database and then I can search the whole thing, or have AI look for interesting patterns in my thoughts.", "context": "", "use_case": "local voice transcription", "switched_from": "", "story": "Ask HN: What are tools you have made for yourself since the advent of AI"}, {"id": 774205921, "item_id": 48387870, "url": "https://news.ycombinator.com/item?id=48387870", "by": "quickthoughts", "date": "2026-06-03", "stance": "neutral", "quote": "I actually built out a fine-tuning whisper pipeline with all local (smaller) models", "excerpt": "I use small models like Gemma to improve transcriptions from ASR models amongst other micro-tasks. I actually built out a fine-tuning whisper pipeline with all local (smaller) models meaning no cloud/big-tech co is able to train/sell my (private) data. Repo is https://github.com/Rebreda/listenr - mainly geared toward Whisper fine-tuning, AMD hardware and local inference", "context": "", "use_case": "local speech recognition fine-tuning", "switched_from": "", "story": "Gemma 4 12B: A unified, encoder-free multimodal model"}, {"id": 773500016, "item_id": 48343751, "url": "https://news.ycombinator.com/item?id=48343751", "by": "sureglymop", "date": "2026-05-31", "stance": "recommend", "quote": "smaller specialised things like whisper, CLIP, SAM, etc.", "excerpt": "What I'm actually excited about in AI are smaller specialised things like whisper, CLIP, SAM, etc. Just stacking them together as little building blocks has been amazing. Great app to hopefully let more people realize that it's not all just huge LLMs that have to run somewhere else!", "context": "", "use_case": "speech recognition building block", "switched_from": "", "story": "Ask HN: What Is the State of App Development in 2026?"}, {"id": 771699058, "item_id": 48231191, "url": "https://news.ycombinator.com/item?id=48231191", "by": "carpo", "date": "2026-05-21", "stance": "neutral", "quote": "I made it a local Electron app with Whisper, ffmpeg", "excerpt": "This is great. I wish I had enough ram for a local model. I just spent the last few weeks writing something very similar, but I made it a local Electron app with Whisper, ffmpeg and I added semantic search and embeddings for chatting with the videos. It talks to Claude for the vision analysis, tagging and video chat. Do you only send one image for yours? I used a customised scene detection algorithm to find multiple ", "context": "", "use_case": "video transcription", "switched_from": "", "story": "Indexing a year of video locally on a 2021 MacBook with Gemma4-31B (50GB swap)"}, {"id": 768243346, "item_id": 48015209, "url": "https://news.ycombinator.com/item?id=48015209", "by": "pncnmnp", "date": "2026-05-04", "stance": "neutral", "quote": "so I channel it all via Whisper.", "excerpt": "The whole setup works on my M2 MacBook Pro with 16 GB RAM. I use Gemma 4B via LiteRT-LM. I've found that LiteRT-LM has a much lower DRAM footprint than Ollama. I've also made tons of optimizations in the code - for eg, you can do quite a bit with a 16k context window for a voice assistant while managing a good footprint, so I keep track of the token usage and then perform an auto-compaction after a while. I use sub-a", "context": "", "use_case": "speech recognition", "switched_from": "", "story": "How OpenAI delivers low-latency voice AI at scale"}, {"id": 767643826, "item_id": 47977739, "url": "https://news.ycombinator.com/item?id=47977739", "by": "nerdsniper", "date": "2026-05-01", "stance": "recommend", "quote": "Whisper models in general seem to do a lot better than most built-into-OS/app models.", "excerpt": "On just the transcription side: WhisprFlow produces much better speech-to-text for long text messaging-by-voice (dictation / transcription) than apple's speech-to-text does. Whisper models in general seem to do a lot better than most built-into-OS/app models. Which is interesting, because there's nothing stopping them from just using Whisper models. I love MacWhisper personally. Also, Gumroad is a fantastic app distr", "context": "", "use_case": "speech-to-text", "switched_from": "", "story": "Apple accidentally left Claude.md files Apple Support app"}, {"id": 756537411, "item_id": 47283588, "url": "https://news.ycombinator.com/item?id=47283588", "by": "moirnz", "date": "2026-03-06", "stance": "neutral", "quote": "Uses openai whisper for speech recognition", "excerpt": "I'm new to this so bear with me. The coder works on Ollama models or Cloude. Take your pick. Can type the problem in or use speech recognition and that's it. tommoirnz/NLAutocoder: Natural Language AutoCoder v1.0 tommoirnz/autocoder-error-cache: Seeded error cache for autocoder. For cloud based ie say Claude it rarely makes mistakes and can retrieve from the internet. For ollama models it has a seeded error_cache you", "context": "", "use_case": "speech recognition", "switched_from": "", "story": "Natural Language AutoCoder Open SOurce"}, {"id": 755817061, "item_id": 47238566, "url": "https://news.ycombinator.com/item?id=47238566", "by": "ali_chherawalla", "date": "2026-03-03", "stance": "neutral", "quote": "Voice is on-device Whisper.", "excerpt": "ofcourse ofcourse, I've documented everything here: https://github.com/alichherawalla/off-grid-mobile-ai/blob/main/docs/ARCHITECTURE.md llama.cpp compiled as a native Android library via the NDK, linked into React Native through a custom JSI bridge. GGUF models loaded straight into memory. On Snapdragon devices we use QNN (Qualcomm Neural Network) for hardware acceleration. OpenCL GPU fallback on everything else. CPU", "context": "built an on-device Android AI app", "use_case": "on-device voice", "switched_from": "", "story": "Show HN: Qwen 3.5 running on a $300 Android phone \u2013 on-device, open source"}, {"id": 751940768, "item_id": 46996298, "url": "https://news.ycombinator.com/item?id=46996298", "by": "srinifromsalem", "date": "2026-02-12", "stance": "recommend", "quote": "Whisper remains the gold standard - you can run it locally with good accuracy across languages.", "excerpt": "For local speech-to-text, Whisper remains the gold standard - you can run it locally with good accuracy across languages. For speech-to-speech, you'd typically chain Whisper with a local TTS model like Coqui TTS or use something like Tortoise TTS for higher quality but slower processing. The key is balancing accuracy, speed, and resource usage based on your specific use case. If you're doing content creation workflow", "context": "", "use_case": "local speech-to-text", "switched_from": "", "story": "Ask HN: What's the current best local/open speech-to-speech setup?"}, {"id": 751599540, "item_id": 46974971, "url": "https://news.ycombinator.com/item?id=46974971", "by": "prradox", "date": "2026-02-11", "stance": "neutral", "quote": "STT via Whisper", "excerpt": "I run multiple Claude Code and OpenCode sessions at the same time \u2014 sometimes on my laptop, sometimes on remote devboxes over SSH. The friction that bugged me most was context switching: I'd be thinking through a problem, pacing around, and then have to sit down, find the right terminal window, and type out what I wanted the agent to do. I kept wishing I could just say it out loud. So I built AgentWire. It's a self-h", "context": "", "use_case": "speech-to-text", "switched_from": "", "story": "Show HN: AgentWire \u2013 Talk to your AI coding agents by voice, across machines"}, {"id": 751499876, "item_id": 46968742, "url": "https://news.ycombinator.com/item?id=46968742", "by": "prradox", "date": "2026-02-10", "stance": "neutral", "quote": "Whisper for STT", "excerpt": "I run several Claude Code and OpenCode sessions at once, often across 2+ devboxes and my local machine. I wanted a way to talk to my Claude Code session without having to be in the window/terminal. I wanted a way to see all my tmux sessions across machines from a single interface. So, I built agentwire-dev. AgentWire is a self-hosted web portal and CLI that sits on top of tmux. It gives you a single dashboard where y", "context": "", "use_case": "speech-to-text", "switched_from": "", "story": "Show HN: One UI to voice-control local and remote AI coding agents via tmux"}, {"id": 750319986, "item_id": 46894999, "url": "https://news.ycombinator.com/item?id=46894999", "by": "d4rkp4ttern", "date": "2026-02-04", "stance": "compare", "quote": "accuracy slightly less than slower Whisper models", "excerpt": "I\u2019m curious about this too. On my M1 Max MacBook I use the Handy app on macOS with Parakeet V3 and I get near instant transcription, accuracy slightly less than slower Whisper models, but that drop is immaterial when talking to CLI coding agents, which is where I find the most use for this. https://github.com/cjpais/Handy", "context": "", "use_case": "transcription", "switched_from": "", "story": "Voxtral Transcribe 2"}, {"id": 747879313, "item_id": 46742457, "url": "https://news.ycombinator.com/item?id=46742457", "by": "dvfjsdhgfv", "date": "2026-01-24", "stance": "recommend", "quote": "Whisper is really good for 1 person speaking.", "excerpt": "I spent a few days on similar scenario without much success (scenario where one person speaks and then their speech is translated, and I want juts the original or both). An API call to GPT4o works quite well (it basically handles both transcription and diarization), but I wanted a local model. Whisper is really good for 1 person speaking. With more people you get repetitions. Qwen and other open multimodal models giv", "context": "", "use_case": "speech-to-text", "switched_from": "", "story": "Ask HN: What's the current best local/open speech-to-speech setup?"}, {"id": 747873680, "item_id": 46742105, "url": "https://news.ycombinator.com/item?id=46742105", "by": "woudsma", "date": "2026-01-24", "stance": "recommend", "quote": "I\u2019m using whisper with superwhisper on my mac.", "excerpt": "I\u2019m using whisper with superwhisper on my mac. I\u2019ve assigned a key on my keyboard, when I press the key it starts listening and when I release it, the text gets copied to the current cursor location. It works pretty well.", "context": "", "use_case": "local dictation on Mac", "switched_from": "", "story": "Ask HN: What's the current best local/open speech-to-speech setup?"}, {"id": 747854353, "item_id": 46740897, "url": "https://news.ycombinator.com/item?id=46740897", "by": "benatkin", "date": "2026-01-23", "stance": "recommend", "quote": "it would make more sense to go straight to Whisper.", "excerpt": "To save a click, it's just a fancy front end for Whisper plus a weaker CPU-only model. It has a demo video that seems impressive, but the speech is careful to sound casual while having no meaningful flaws that would cause it to mess up. If you want to make a speech to speech tool, which is what this post asks about, it would make more sense to go straight to Whisper.", "context": "", "use_case": "speech-to-speech tooling", "switched_from": "", "story": "Ask HN: What's the current best local/open speech-to-speech setup?"}, {"id": 747832992, "item_id": 46739562, "url": "https://news.ycombinator.com/item?id=46739562", "by": "BiraIgnacio", "date": "2026-01-23", "stance": "compare", "quote": "I'll probably go with whisper (with whisper.cpp) but has anyone compared it to vosk models?", "excerpt": "oh I've been looking into whisper and vosk in the last few days. I'll probably go with whisper (with whisper.cpp) but has anyone compared it to vosk models?", "context": "", "use_case": "speech-to-text", "switched_from": "", "story": "Ask HN: What's the current best local/open speech-to-speech setup?"}, {"id": 747832704, "item_id": 46739544, "url": "https://news.ycombinator.com/item?id=46739544", "by": "cbrews", "date": "2026-01-23", "stance": "neutral", "quote": "streaming Speech-to-Text (using Whisper)", "excerpt": "Thanks for sharing this! I'm going to put this on my list to play around with. I'm not really an expert in this tech, I come from the audio background, but recently was playing around with streaming Speech-to-Text (using Whisper) / Text-to-Speech (using Kokoro at the time) on a local machine. The most challenging part in my build was tuning the inference batch sizing here. I was able to get it working well for Speech", "context": "audio background", "use_case": "streaming speech-to-text", "switched_from": "", "story": "Ask HN: What's the current best local/open speech-to-speech setup?"}, {"id": 747812773, "item_id": 46738298, "url": "https://news.ycombinator.com/item?id=46738298", "by": "jauntywundrkind", "date": "2026-01-23", "stance": "recommend", "quote": "those are amazing models", "excerpt": "It was a little annoying getting old qt5 tools installed but I really enjoyed using dsnote / Speech Note. Huge model selection for my amd gpu. Good tool. I haven't done enough specific studying yet to give you suggestions for which model to go with. WhisperFlow is very popular. Kyutai some very interesting work always. Their delayed streams work is bleeding edge & sounds very promising especially for low latency. Not", "context": "", "use_case": "speech recognition", "switched_from": "", "story": "Ask HN: What's the current best local/open speech-to-speech setup?"}, {"id": 744973633, "item_id": 46560852, "url": "https://news.ycombinator.com/item?id=46560852", "by": "delaminator", "date": "2026-01-09", "stance": "recommend", "quote": "My favourite use is Whisper. I voice->text almost all of my typing now.", "excerpt": "I have a 3090 24Gb Twin Xenon 64Gb RAM sat on a machine in our server room. I do local AI with Qwen, Whisper and another I can't remember right now. These are all QWEN: We do AI Invoice OCR - PDF -> Image -> Excel. Works much better than other solutions because it has invoice context so it looks for particular data to extract and ignores others. Why local? I proved it worked, no need to send our data outside for proc", "context": "", "use_case": "voice dictation", "switched_from": "", "story": "Ask HN: Who's running local AI workstations in 2026?"}, {"id": 744520368, "item_id": 46532523, "url": "https://news.ycombinator.com/item?id=46532523", "by": "delaminator", "date": "2026-01-07", "stance": "recommend", "quote": "Whisper works great, even the medium model is pretty good.", "excerpt": "Whisper works great, even the medium model is pretty good. But I use the 3Gb all day every day. I built a personal voice agent https://github.com/lawless-m/TheHand", "context": "", "use_case": "local speech recognition", "switched_from": "", "story": "Dell admits consumers don't care about AI PCs"}, {"id": 744356707, "item_id": 46522294, "url": "https://news.ycombinator.com/item?id=46522294", "by": "ec109685", "date": "2026-01-06", "stance": "neutral", "quote": "I use Prompt, Ever Terminal, Whisper, EC2 and Claude Code.", "excerpt": "I use Prompt, Ever Terminal, Whisper, EC2 and Claude Code. I can build anything with it. Having Claude on top of a terraform repo lets me fully control my infra. Claude is so good at AWS and terraform, and it even found a $3k monthly accidental spend I had running (also sent a refund request to hopefully get some credit back). Also have a Claude driven CI workflow in GitHub to help keep everything on track. Having fu", "context": "", "use_case": "voice input", "switched_from": "", "story": "Stop Doom Scrolling, Start Doom Coding: Build via the terminal from your phone"}, {"id": 743589877, "item_id": 46474367, "url": "https://news.ycombinator.com/item?id=46474367", "by": "delaminator", "date": "2026-01-03", "stance": "neutral", "quote": "I also run a Whisper server and have built a voice-keyboard program.", "excerpt": "I do a lot of AI coding Receipts: https://github.com/lawless-m?tab=repositories I started in Cursor - the tab completion is superb. As an assistant to the coder it is incredible. But then I started to lean on Sonnet more and more. I expressed my ideas and they came alive. As I got better at prompting I can code at the speed of thought. So I switched to Claude Code entirely. I\u2019m a Max 200 customer now. On Windows I st", "context": "", "use_case": "voice-to-text dictation", "switched_from": "", "story": "Ask HN: Do you prefer AI coding in an IDE or CLI? And why?"}, {"id": 734901824, "item_id": 45931364, "url": "https://news.ycombinator.com/item?id=45931364", "by": "lxe", "date": "2025-11-14", "stance": "recommend", "quote": "I much prefer whisper after playing with both. Maybe I'm not configuring Parakeet correctly, But the transcription that comes out of Whisper is usually pretty much spot on.", "excerpt": "I've done something similar for Linux and Mac. I originally used Whisper and then switched to Parakeet. I much prefer whisper after playing with both. Maybe I'm not configuring Parakeet correctly, But the transcription that comes out of Whisper is usually pretty much spot on. It automatically removes all the \"ooms\" and all the \"ahs\" and it's just way more natural, in my opinion. I'm using Whisper.CPP with CUDA accele", "context": "built dictation tools for Linux and Mac", "use_case": "voice dictation", "switched_from": "Parakeet", "story": "Show HN: Chirp \u2013 Local Windows dictation with ParakeetV3 no executable required"}, {"id": 734830208, "item_id": 45926888, "url": "https://news.ycombinator.com/item?id=45926888", "by": "geerlingguy", "date": "2025-11-14", "stance": "recommend", "quote": "Whisper has radically improved CC on my video content.", "excerpt": "I like the conclusion; like for me, Whisper has radically improved CC on my video content. I used to spend a few hours translating my scripts into CCs, and tooling was poor. Now I run it through whisper in a couple minutes, give one quick pass to correct a few small hallucinations and misspellings, and I'm done. There are big wins in AI. But those don't pump the bubble once they're solved. And the thing that made Whi", "context": "creates video content", "use_case": "video captions transcription", "switched_from": "", "story": "AGI fantasy is a blocker to actual engineering"}]}