{"slug": "nvidia-parakeet", "name": "NVIDIA Parakeet", "category": "ai-models", "category_name": "AI models", "description": "NVIDIA's family of open-weight speech-to-text models for fast automatic speech recognition.", "recommenders": 62, "commenters": 101, "warners": 7, "score": 55.042, "rising": -1.36, "first": "2025-11-06", "latest": "2026-10-29", "why": "Commenters praise it as fast, accurate on-device speech-to-text, often preferred over Whisper and used daily for dictation through apps like Handy and Hex.", "why_is_quote": false, "has_verdict": true, "sentiment": {"recommend": 69, "neutral": 30, "compare": 7, "warn": 8}, "spark": [4, 3, 4, 12, 8, 7, 1, 5, 11, 2, 4, 8], "listed": true, "loved": ["Speed", "Daily dictation"], "caveat": "Multilingual accuracy", "short_quote": "I haven\u2019t found an ASR/STT/transcription setup that beats Parakeet V3 on the speed/accuracy tradeoff spectrum", "kind": "", "icon": false, "domain": "narcotic.sh", "github_repo": null, "official_url": "https://huggingface.co/nvidia/parakeet-tdt-0.6b-v3", "pricing": "free", "aliases": ["Parakeet", "Parakeet V3", "Parakeet v3", "NVIDIA Parakeet", "Parakeet TDT v3", "Parakeet TDT 0.6B", "NVIDIA Parakeet TDT", "Parakeet 3", "Parakeet TDT 0.6B v2", "Parakeet TDT", "Parakeet TDT 0.6B v3", "Nvidia Parakeet", "NVIDIA Parakeet V3", "Nvidia Parakeet V3"], "verdict": [{"text": "Commenters praise it as fast, accurate on-device speech-to-text", "cites": [{"id": 782406273, "item_id": 48900392, "url": "https://news.ycombinator.com/item?id=48900392", "by": "hamza_q_", "date": "2026-07-13", "version": null, "stance": "recommend", "quote": "sped up Parakeet V2 and V3 to 320x and 282x faster than real time", "excerpt": "Recently contributed a patch to FluidAudio that sped up Parakeet V2 and V3 to 320x and 282x faster than real time, respectively: https://github.com/FluidInference/FluidAudio/pull/507 That means one hour of audio transcribed in 11.25 and 12.75 seconds. The Inscribe post doesn't give a speed factor for SpeechAnalyzer. However, this Argmax blog post reports 70: https://www.argmaxinc.com/blog/apple-and-argmax Based on th", "context": "", "use_case": "Speech-to-text model", "switched_from": "", "story": "Apple's new SpeechAnalyzer API, benchmarked against Whisper and its predecessor"}, {"id": 777394448, "item_id": 48587153, "url": "https://news.ycombinator.com/item?id=48587153", "by": "rob", "date": "2026-06-18", "version": null, "stance": "recommend", "quote": "was completely surprised at just how fast it was for being on-device!", "excerpt": "Yep, here's the v2 and v3: https://huggingface.co/nvidia/parakeet-tdt-0.6b-v2 https://huggingface.co/nvidia/parakeet-tdt-0.6b-v3 It's almost instant on my new M5 Max w/ 36GB of memory, but I used both with Handy on my previous 2019 Intel Mac w/ 16GB memory and was completely surprised at just how fast it was for being on-device! Not instant, but only a couple seconds.", "context": "", "use_case": "on-device speech-to-text", "switched_from": "", "story": "DeepSeek Introduces Vision"}, {"id": 802705664, "item_id": 50169104, "url": "https://news.ycombinator.com/item?id=50169104", "by": "7777777phil", "date": "2026-10-20", "version": null, "stance": "recommend", "quote": "been pretty impressed with the local performance", "excerpt": "I have been using different Parakeet models for STT over the past months and been pretty impressed with the local performance.", "context": "", "use_case": "local speech-to-text", "switched_from": "", "story": "Voxtral 3"}, {"id": 780447008, "item_id": 48777938, "url": "https://news.ycombinator.com/item?id=48777938", "by": "randomblock1", "date": "2026-07-03", "version": "3", "stance": "recommend", "quote": "Parakeet TDT v3 is half the size, way faster, and lower WER", "excerpt": "No, there are quite a few models which are smaller, more accurate, and faster. For example Parakeet TDT v3 is half the size, way faster, and lower WER. There's also Voxstral, which is much larger but also even more accurate. But the ecosystem isn't as mature, so Whisper is still a valid option, even now. For example Parakeet uses Nemotron framework (made by Nvdia), normally you need CUDA, so you need to use an ONNX v", "context": "", "use_case": "", "switched_from": "", "story": "Jamesob's guide to running SOTA LLMs locally"}]}, {"text": ", often preferred over Whisper", "cites": [{"id": 782324128, "item_id": 48895258, "url": "https://news.ycombinator.com/item?id=48895258", "by": "coder543", "date": "2026-07-13", "version": null, "stance": "recommend", "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": 782335713, "item_id": 48895982, "url": "https://news.ycombinator.com/item?id=48895982", "by": "trencedamp", "date": "2026-07-13", "version": null, "stance": "recommend", "quote": "it's perfect with parakeet.", "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": "Whisper", "story": "Apple's new SpeechAnalyzer API, benchmarked against Whisper and its predecessor"}, {"id": 779296673, "item_id": 48706042, "url": "https://news.ycombinator.com/item?id=48706042", "by": "taffydavid", "date": "2026-06-28", "version": "3", "stance": "recommend", "quote": "you can use parakeet v3 which I find way better than whisper", "excerpt": "I love on-device transcription apps, but this looks like yet another AI generated whisper wrapper. I'm put off by the readme which is clearly LLM generated, with emdash all over the place and bits that smell of user questions turned into talking points. > Why not the Fn / key? It's the obvious one-finger choice, but on macOS the Fn key emits no real keypress \u2014 only a hidden hardware flag \u2014 so And > macOS permissions ", "context": "", "use_case": "speech transcription", "switched_from": "", "story": "Yap \u2013 free offline voice dictation for Mac/Windows/Linux (Wispr Flow alt)"}, {"id": 791637873, "item_id": 49477367, "url": "https://news.ycombinator.com/item?id=49477367", "by": "zackify", "date": "2026-08-28", "version": null, "stance": "recommend", "quote": "Also whisper is pretty outdated vs parakeet", "excerpt": "https://voxtype.io/ https://tryvoiceink.com/ There's so many of these, at this point I've seen 10 clones make the front page each time as if there never existed local only options before. Also whisper is pretty outdated vs parakeet", "context": "", "use_case": "", "switched_from": "", "story": "Show HN: FnScribe \u2013 Open-source, offline dictation for macOS"}]}, {"text": " and used daily for dictation through apps like Handy and Hex.", "cites": [{"id": 782354816, "item_id": 48897176, "url": "https://news.ycombinator.com/item?id=48897176", "by": "sherlock-holmes", "date": "2026-07-13", "version": null, "stance": "recommend", "quote": "i am using parakeet with https://handy.computer daily and it's amazing.", "excerpt": "i wish they had benchmarked it against parakeet-unified-en-0.6b and cohere-transcribe-03-2026. i am using parakeet with https://handy.computer daily and it's amazing.", "context": "", "use_case": "Speech-to-text dictation", "switched_from": "", "story": "Apple's new SpeechAnalyzer API, benchmarked against Whisper and its predecessor"}, {"id": 800221152, "item_id": 50013822, "url": "https://news.ycombinator.com/item?id=50013822", "by": "andai", "date": "2026-10-08", "version": null, "stance": "recommend", "quote": "Parakeet is great I use it everyday", "excerpt": "Technical knowhow? Handy is a gui right? (Also Parakeet is great I use it everyday)", "context": "", "use_case": "speech-to-text dictation", "switched_from": "", "story": "Whistle: Speech to Text in 16.9 MB"}, {"id": 783497233, "item_id": 48968577, "url": "https://news.ycombinator.com/item?id=48968577", "by": "d4rkp4ttern", "date": "2026-07-19", "version": "3", "stance": "recommend", "quote": "near instant transcription with Parakeet V3", "excerpt": "I use the Hex[1] app on MacOs for near instant transcription with Parakeet V3. This is how I speak to Code agents at least 80% of the time. The idea is - I tend to be lazy if I need to type lots of details, so speaking lets me get into details that I otherwise wouldn\u2019t and this helps give more useful context to the agent. This often tends to be an unstructured brain dump so I sometimes ask the agent to repeat back wh", "context": "", "use_case": "speech transcription", "switched_from": "", "story": "Speech Recognition and TTS in less than 500kb"}, {"id": 782379585, "item_id": 48898724, "url": "https://news.ycombinator.com/item?id=48898724", "by": "bel8", "date": "2026-07-13", "version": "3", "stance": "recommend", "quote": "It's  great for me when configured to use Parakeet v3.", "excerpt": "Obligatory mention of Handy which is open-source and cross-platform: https://github.com/cjpais/Handy It's great for me when configured to use Parakeet v3.", "context": "", "use_case": "speech transcription", "switched_from": "", "story": "Apple's new SpeechAnalyzer API, benchmarked against Whisper and its predecessor"}]}, {"text": " The most common caveat is weaker multilingual handling, such as misdetected languages and missed non-English terms", "cites": [{"id": 782482145, "item_id": 48905134, "url": "https://news.ycombinator.com/item?id=48905134", "by": "rofrol", "date": "2026-07-14", "version": "3", "stance": "warn", "quote": "I speak Polish and it detects Russian.", "excerpt": "App works great. Parkeeter TDT V3 not so much. I speak Polish and it detects Russian. Also there is no possibility to force language: > unfortuntately parakeet-v3 model doesn;t recieve or output language id https://github.com/NVIDIA-NeMo/Speech/issues/14799#issuecomment-3634425420 Also this: > If you are using Parakeet for English only then you should be using V2. V3 is for several languages and is worse at English o", "context": "", "use_case": "Speech recognition", "switched_from": "", "story": "Apple's new SpeechAnalyzer API, benchmarked against Whisper and its predecessor"}, {"id": 791562720, "item_id": 49472670, "url": "https://news.ycombinator.com/item?id=49472670", "by": "Lucasoato", "date": "2026-08-27", "version": "3", "stance": "recommend", "quote": "I love it but it misses the business specific words when in different language.", "excerpt": "I love it but it misses the business specific words when in different language. Sometimes it stretches them out to fit unrelated words in the language of the rest of the conversation. I miss its speed though.", "context": "", "use_case": "speech transcription", "switched_from": "", "story": "Gemini-3.5-Transcribe"}, {"id": 797773937, "item_id": 49860871, "url": "https://news.ycombinator.com/item?id=49860871", "by": "gunalx", "date": "2026-09-26", "version": null, "stance": "recommend", "quote": "parakeet for speed", "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": "fast speech recognition", "switched_from": "", "story": "Ask HN: What would be the best local ASR model"}]}, {"text": ", and a few users still find Whisper more accurate for their needs.", "cites": [{"id": 791599409, "item_id": 49474963, "url": "https://news.ycombinator.com/item?id=49474963", "by": "jwr", "date": "2026-08-28", "version": null, "stance": "warn", "quote": "I keep hearing that Nvidia Parakeet models are better, but they just don't work as well for me", "excerpt": "It's interesting that the results can be so different depending on the person, the use case and even the microphone used. I use dictation a lot, so I try to stay up to date with the latest models as much as I can. So far, for my needs, nothing could beat Whisper Large v3. I keep hearing that Nvidia Parakeet models are better, but they just don't work as well for me, even though they are unquestionably faster. Things ", "context": "", "use_case": "dictation", "switched_from": "", "story": "Gemini-3.5-Transcribe"}]}], "praise": [{"label": "Speed", "summary": "Commenters call it near-instant, with one reporting 282x to 320x faster than real time in an optimized runtime.", "quotes": [{"id": 782406273, "item_id": 48900392, "url": "https://news.ycombinator.com/item?id=48900392", "by": "hamza_q_", "date": "2026-07-13", "version": null, "stance": "recommend", "quote": "sped up Parakeet V2 and V3 to 320x and 282x faster than real time", "excerpt": "Recently contributed a patch to FluidAudio that sped up Parakeet V2 and V3 to 320x and 282x faster than real time, respectively: https://github.com/FluidInference/FluidAudio/pull/507 That means one hour of audio transcribed in 11.25 and 12.75 seconds. The Inscribe post doesn't give a speed factor for SpeechAnalyzer. However, this Argmax blog post reports 70: https://www.argmaxinc.com/blog/apple-and-argmax Based on th", "context": "", "use_case": "Speech-to-text model", "switched_from": "", "story": "Apple's new SpeechAnalyzer API, benchmarked against Whisper and its predecessor"}, {"id": 783497233, "item_id": 48968577, "url": "https://news.ycombinator.com/item?id=48968577", "by": "d4rkp4ttern", "date": "2026-07-19", "version": "3", "stance": "recommend", "quote": "near instant transcription with Parakeet V3", "excerpt": "I use the Hex[1] app on MacOs for near instant transcription with Parakeet V3. This is how I speak to Code agents at least 80% of the time. The idea is - I tend to be lazy if I need to type lots of details, so speaking lets me get into details that I otherwise wouldn\u2019t and this helps give more useful context to the agent. This often tends to be an unstructured brain dump so I sometimes ask the agent to repeat back wh", "context": "", "use_case": "speech transcription", "switched_from": "", "story": "Speech Recognition and TTS in less than 500kb"}, {"id": 777394448, "item_id": 48587153, "url": "https://news.ycombinator.com/item?id=48587153", "by": "rob", "date": "2026-06-18", "version": null, "stance": "recommend", "quote": "was completely surprised at just how fast it was for being on-device!", "excerpt": "Yep, here's the v2 and v3: https://huggingface.co/nvidia/parakeet-tdt-0.6b-v2 https://huggingface.co/nvidia/parakeet-tdt-0.6b-v3 It's almost instant on my new M5 Max w/ 36GB of memory, but I used both with Handy on my previous 2019 Intel Mac w/ 16GB memory and was completely surprised at just how fast it was for being on-device! Not instant, but only a couple seconds.", "context": "", "use_case": "on-device speech-to-text", "switched_from": "", "story": "DeepSeek Introduces Vision"}]}, {"label": "Daily dictation", "summary": "Many use it every day for dictation through apps like Handy, Hex and Superwhisper.", "quotes": [{"id": 782354816, "item_id": 48897176, "url": "https://news.ycombinator.com/item?id=48897176", "by": "sherlock-holmes", "date": "2026-07-13", "version": null, "stance": "recommend", "quote": "i am using parakeet with https://handy.computer daily and it's amazing.", "excerpt": "i wish they had benchmarked it against parakeet-unified-en-0.6b and cohere-transcribe-03-2026. i am using parakeet with https://handy.computer daily and it's amazing.", "context": "", "use_case": "Speech-to-text dictation", "switched_from": "", "story": "Apple's new SpeechAnalyzer API, benchmarked against Whisper and its predecessor"}, {"id": 782364176, "item_id": 48897761, "url": "https://news.ycombinator.com/item?id=48897761", "by": "dinfinity", "date": "2026-07-13", "version": "3", "stance": "recommend", "quote": "I use Parakeet V3 via this tool and it is actually quite reliable for me (in English)", "excerpt": "I use Parakeet V3 via this tool and it is actually quite reliable for me (in English): https://github.com/cjpais/Handy", "context": "", "use_case": "speech transcription", "switched_from": "", "story": "Apple's new SpeechAnalyzer API, benchmarked against Whisper and its predecessor"}, {"id": 800150513, "item_id": 50009407, "url": "https://news.ycombinator.com/item?id=50009407", "by": "cgbur", "date": "2026-10-08", "version": null, "stance": "recommend", "quote": "I use https://github.com/cjpais/Handy on Parakeet streaming", "excerpt": "For essentially infinite and fast dictation I use https://github.com/cjpais/Handy on Parakeet streaming (cohere is far better, but slower and has a token output limit so you cant ramble for many minutes). And then just do a cleanup pass with a cheap LLM, it will in my experience, do far better than trying to voice control to go edit a sentence or change words. I just weave instructions into my writing. I understand t", "context": "", "use_case": "streaming speech recognition", "switched_from": "", "story": "Whistle: Speech to Text in 16.9 MB"}]}, {"label": "Beats Whisper", "summary": "Several commenters say it is more accurate and faster than Whisper, with one calling it the gold standard.", "quotes": [{"id": 782324128, "item_id": 48895258, "url": "https://news.ycombinator.com/item?id=48895258", "by": "coder543", "date": "2026-07-13", "version": null, "stance": "recommend", "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": 782335713, "item_id": 48895982, "url": "https://news.ycombinator.com/item?id=48895982", "by": "trencedamp", "date": "2026-07-13", "version": null, "stance": "recommend", "quote": "it's perfect with parakeet.", "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": "Whisper", "story": "Apple's new SpeechAnalyzer API, benchmarked against Whisper and its predecessor"}, {"id": 800162656, "item_id": 50010166, "url": "https://news.ycombinator.com/item?id=50010166", "by": "theturtletalks", "date": "2026-10-08", "version": null, "stance": "recommend", "quote": "Parakeet is the gold standard.", "excerpt": "Parakeet is the gold standard. With models like moonshine and koroko (TTS model), it\u2019s more about embedding the model in the application itself. If you\u2019re using Parakeet, embedding it in the application is not feasible. I use parakeet with superwhisper, and I\u2019m making another app that has SST and TTS built in, and I want to use my downloaded parakeet model, but it seems there\u2019s so many different implementations from ", "context": "building an app with STT and TTS", "use_case": "speech-to-text", "switched_from": "", "story": "Whistle: Speech to Text in 16.9 MB"}]}, {"label": "Runs locally", "summary": "Users are impressed that it runs well on-device, including on older Macs and an iPhone.", "quotes": [{"id": 802705664, "item_id": 50169104, "url": "https://news.ycombinator.com/item?id=50169104", "by": "7777777phil", "date": "2026-10-20", "version": null, "stance": "recommend", "quote": "been pretty impressed with the local performance", "excerpt": "I have been using different Parakeet models for STT over the past months and been pretty impressed with the local performance.", "context": "", "use_case": "local speech-to-text", "switched_from": "", "story": "Voxtral 3"}, {"id": 782392945, "item_id": 48899559, "url": "https://news.ycombinator.com/item?id=48899559", "by": "cootsnuck", "date": "2026-07-13", "version": "3", "stance": "recommend", "quote": "Nvidia's Parakeet v3 is good enough for my own just local transcription most of the time.", "excerpt": "For just desktop, I use Handy (open source): https://handy.computer/ And yea, Nvidia's Parakeet v3 is good enough for my own just local transcription most of the time. When I need local transcription to be more reliable and I don't have the energy to proof read a long ramble, I still often just pop open chatGPT, dictate, cut, paste. But we're pretty much already to the point where local transcription models can repla", "context": "", "use_case": "local transcription", "switched_from": "", "story": "Apple's new SpeechAnalyzer API, benchmarked against Whisper and its predecessor"}, {"id": 782324128, "item_id": 48895258, "url": "https://news.ycombinator.com/item?id=48895258", "by": "coder543", "date": "2026-07-13", "version": null, "stance": "recommend", "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"}]}, {"label": "Talking to agents", "summary": "Its speed and accuracy suit dictating prompts to AI coding agents, where small errors matter less.", "quotes": [{"id": 800788304, "item_id": 50049269, "url": "https://news.ycombinator.com/item?id=50049269", "by": "d4rkp4ttern", "date": "2026-10-11", "version": "3", "stance": "recommend", "quote": "dictation was solved a while back with Parakeet V3", "excerpt": "On MacOS, dictation was solved a while back with Parakeet V3, and I haven't seen anything that has a better speed/accuracy tradeoff when talking to AIs, where slight inaccuracies are ok since they can \"read between the lines\". I got annoyed by various UI and other things in tools like VoiceInk, Handy etc and Claude-Coded my own CLI dictation tool called VoxType that uses parakeet-mlx: https://pchalasani.github.io/cla", "context": "", "use_case": "speech recognition", "switched_from": "", "story": "Show HN: BetterWispr \u2013 Free, open-source dictation for Mac"}, {"id": 783497233, "item_id": 48968577, "url": "https://news.ycombinator.com/item?id=48968577", "by": "d4rkp4ttern", "date": "2026-07-19", "version": "3", "stance": "recommend", "quote": "near instant transcription with Parakeet V3", "excerpt": "I use the Hex[1] app on MacOs for near instant transcription with Parakeet V3. This is how I speak to Code agents at least 80% of the time. The idea is - I tend to be lazy if I need to type lots of details, so speaking lets me get into details that I otherwise wouldn\u2019t and this helps give more useful context to the agent. This often tends to be an unstructured brain dump so I sometimes ask the agent to repeat back wh", "context": "", "use_case": "speech transcription", "switched_from": "", "story": "Speech Recognition and TTS in less than 500kb"}]}, {"label": "Voice pipelines", "summary": "Commenters use it as the speech-to-text stage in local voice assistants and custom devices.", "quotes": [{"id": 800362640, "item_id": 50022665, "url": "https://news.ycombinator.com/item?id=50022665", "by": "cco", "date": "2026-10-09", "version": null, "stance": "recommend", "quote": "parakeet for STT", "excerpt": "Yeah maybe I'll play around with different precision, I did see improvement when I went from fp4 to full fidelity though. This is my setup for Home Assistant voice, parakeet for STT, GLM 5.3 to generate responses, and Kokoro for TTS. I use local routing for simple commands like turn on the lights, but if I want to ask questions or something, it gets routed to GLM 5.3. It is basically free, half a cent per turn maybe,", "context": "", "use_case": "speech-to-text for Home Assistant voice", "switched_from": "", "story": "Whistle: Speech to Text in 16.9 MB"}, {"id": 781186609, "item_id": 48824163, "url": "https://news.ycombinator.com/item?id=48824163", "by": "janpmz", "date": "2026-07-07", "version": null, "stance": "recommend", "quote": "Parakeet for STT", "excerpt": "Both Text-to-Speech and Speech-to-Text now have local models that are good enough to get the job done. Kokoro for TTS, Parakeet for STT and Fluid-1 for text formatting (I use it with FluidVoice). I hope this is a trend that continues for other applications.", "context": "", "use_case": "local speech-to-text", "switched_from": "", "story": "Local, CPU-Friendly, High-Quality TTS (Text-to-Speech) with Kokoro"}, {"id": 800242256, "item_id": 50015141, "url": "https://news.ycombinator.com/item?id=50015141", "by": "nl", "date": "2026-10-09", "version": null, "stance": "recommend", "quote": "offloads transcription to Parakeet running in the browser connected to the ESP32", "excerpt": "I recently built a 3D printed ESP32 based transcription device[1] that offloads transcription to Parakeet running in the browser connected to the ESP32. I haven't done my own measurements of error rate but it'd be interesting to see how well this runs on the ESP32 itself. In my case I'm already using one core to run DSP for a beamforming mic array (which works amazingly well for noise cancellation!) so I don't have h", "context": "builds ESP32 transcription device", "use_case": "speech transcription", "switched_from": "", "story": "Whistle: Speech to Text in 16.9 MB"}]}], "caveats": [{"label": "Multilingual accuracy", "summary": "V3 can misdetect languages, cannot be forced to one language, and misses business terms in other languages.", "quotes": [{"id": 782482145, "item_id": 48905134, "url": "https://news.ycombinator.com/item?id=48905134", "by": "rofrol", "date": "2026-07-14", "version": "3", "stance": "warn", "quote": "I speak Polish and it detects Russian.", "excerpt": "App works great. Parkeeter TDT V3 not so much. I speak Polish and it detects Russian. Also there is no possibility to force language: > unfortuntately parakeet-v3 model doesn;t recieve or output language id https://github.com/NVIDIA-NeMo/Speech/issues/14799#issuecomment-3634425420 Also this: > If you are using Parakeet for English only then you should be using V2. V3 is for several languages and is worse at English o", "context": "", "use_case": "Speech recognition", "switched_from": "", "story": "Apple's new SpeechAnalyzer API, benchmarked against Whisper and its predecessor"}, {"id": 791562720, "item_id": 49472670, "url": "https://news.ycombinator.com/item?id=49472670", "by": "Lucasoato", "date": "2026-08-27", "version": "3", "stance": "recommend", "quote": "I love it but it misses the business specific words when in different language.", "excerpt": "I love it but it misses the business specific words when in different language. Sometimes it stretches them out to fit unrelated words in the language of the rest of the conversation. I miss its speed though.", "context": "", "use_case": "speech transcription", "switched_from": "", "story": "Gemini-3.5-Transcribe"}, {"id": 797773937, "item_id": 49860871, "url": "https://news.ycombinator.com/item?id=49860871", "by": "gunalx", "date": "2026-09-26", "version": null, "stance": "recommend", "quote": "parakeet for speed", "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": "fast speech recognition", "switched_from": "", "story": "Ask HN: What would be the best local ASR model"}]}, {"label": "Varies by user", "summary": "One heavy dictation user finds Whisper Large v3 more accurate for their needs despite Parakeet's speed.", "quotes": [{"id": 791599409, "item_id": 49474963, "url": "https://news.ycombinator.com/item?id=49474963", "by": "jwr", "date": "2026-08-28", "version": null, "stance": "warn", "quote": "I keep hearing that Nvidia Parakeet models are better, but they just don't work as well for me", "excerpt": "It's interesting that the results can be so different depending on the person, the use case and even the microphone used. I use dictation a lot, so I try to stay up to date with the latest models as much as I can. So far, for my needs, nothing could beat Whisper Large v3. I keep hearing that Nvidia Parakeet models are better, but they just don't work as well for me, even though they are unquestionably faster. Things ", "context": "", "use_case": "dictation", "switched_from": "", "story": "Gemini-3.5-Transcribe"}]}, {"label": "Ecosystem friction", "summary": "It normally needs CUDA, requires ONNX on AMD, and its many implementations make embedding it in apps hard.", "quotes": [{"id": 780447008, "item_id": 48777938, "url": "https://news.ycombinator.com/item?id=48777938", "by": "randomblock1", "date": "2026-07-03", "version": "3", "stance": "recommend", "quote": "Parakeet TDT v3 is half the size, way faster, and lower WER", "excerpt": "No, there are quite a few models which are smaller, more accurate, and faster. For example Parakeet TDT v3 is half the size, way faster, and lower WER. There's also Voxstral, which is much larger but also even more accurate. But the ecosystem isn't as mature, so Whisper is still a valid option, even now. For example Parakeet uses Nemotron framework (made by Nvdia), normally you need CUDA, so you need to use an ONNX v", "context": "", "use_case": "", "switched_from": "", "story": "Jamesob's guide to running SOTA LLMs locally"}, {"id": 800162656, "item_id": 50010166, "url": "https://news.ycombinator.com/item?id=50010166", "by": "theturtletalks", "date": "2026-10-08", "version": null, "stance": "recommend", "quote": "Parakeet is the gold standard.", "excerpt": "Parakeet is the gold standard. With models like moonshine and koroko (TTS model), it\u2019s more about embedding the model in the application itself. If you\u2019re using Parakeet, embedding it in the application is not feasible. I use parakeet with superwhisper, and I\u2019m making another app that has SST and TTS built in, and I want to use my downloaded parakeet model, but it seems there\u2019s so many different implementations from ", "context": "building an app with STT and TTS", "use_case": "speech-to-text", "switched_from": "", "story": "Whistle: Speech to Text in 16.9 MB"}]}, {"label": "Digit output", "summary": "It prints spoken numbers as digits, which one developer says breaks merging of transcription windows.", "quotes": [{"id": 803130848, "item_id": 50195678, "url": "https://news.ycombinator.com/item?id=50195678", "by": "marshalla", "date": "2026-10-22", "version": null, "stance": "recommend", "quote": "Parakeet is pretty amazing, though", "excerpt": "\"I've even had it transcribe words before I've finished saying them! \" lol idk if this part is actually a good thing. It's great when you always speak in predictable sandwiches. I do think Parakeet is pretty amazing, though, so some of these product announcements have limited interest to me, especially since it's cloud and costs per hour, even if it's a low cost. Parakeet is trained on text that has had ITN done on i", "context": "", "use_case": "Speech-to-text", "switched_from": "", "story": "Voxtral 3"}]}], "who": [{"by": "nl", "context": "builds ESP32 transcription device", "url": "https://news.ycombinator.com/item?id=50015141", "date": "2026-10-09"}, {"by": "theturtletalks", "context": "building an app with STT and TTS", "url": "https://news.ycombinator.com/item?id=50010166", "date": "2026-10-08"}, {"by": "zenapollo", "context": "low volume user", "url": "https://news.ycombinator.com/item?id=49811751", "date": "2026-09-23"}, {"by": "Charlieholtz", "context": "creator of Conductor", "url": "https://news.ycombinator.com/item?id=48416150", "date": "2026-06-05"}, {"by": "d4rkp4ttern", "context": "macOS user", "url": "https://news.ycombinator.com/item?id=47948352", "date": "2026-04-29"}, {"by": "Dumbledumb", "context": "M1 Air user", "url": "https://news.ycombinator.com/item?id=47371330", "date": "2026-03-13"}, {"by": "andai", "context": "Runs speech-to-text on CPU of a ten-year-old laptop", "url": "https://news.ycombinator.com/item?id=45849704", "date": "2025-11-07"}], "tags": ["for local dictation", "for Mac users", "for talking to coding agents", "for English transcription", "for home voice assistants"], "use_cases": ["speech-to-text", "speech transcription", "speech recognition", "speech recognition model", "speech-to-text model", "local speech-to-text"], "compared": [{"slug": "whisper", "name": "Whisper", "co": 31, "switched_from": 3, "switched_to": 2}, {"slug": "handy", "name": "Handy", "co": 35, "switched_from": 0, "switched_to": 0}, {"slug": "voiceink", "name": "VoiceInk", "co": 10, "switched_from": 0, "switched_to": 0}, {"slug": "hex", "name": "Hex", "co": 8, "switched_from": 0, "switched_to": 0}, {"slug": "voxtral", "name": "Voxtral", "co": 5, "switched_from": 1, "switched_to": 0}, {"slug": "claude-code", "name": "Claude Code", "co": 7, "switched_from": 0, "switched_to": 0}, {"slug": "kokoro", "name": "Kokoro", "co": 7, "switched_from": 0, "switched_to": 0}, {"slug": "macwhisper", "name": "MacWhisper", "co": 7, "switched_from": 0, "switched_to": 0}], "makers": [{"id": 787739808, "item_id": 49233738, "url": "https://news.ycombinator.com/item?id=49233738", "by": "hamza_q_", "date": "2026-08-09", "version": null, "stance": "recommend", "quote": "NVIDIA Parakeet running locally in the browser", "excerpt": "NVIDIA Parakeet running locally in the browser. Implemented in raw WebGPU compute shaders & SIMD WebAssembly audio frontend: https://parakeet.narcotic.sh The aim is for it to be a high-performance implementation. Can currently transcribe 1 hour of audio in 20 seconds on an M5 Mac.", "context": "", "use_case": "Local speech-to-text in browser", "switched_from": "", "story": "Ask HN: What are you working on? (August 2026)"}, {"id": 785280929, "item_id": 49080058, "url": "https://news.ycombinator.com/item?id=49080058", "by": "aghilmort", "date": "2026-07-27", "version": null, "stance": "neutral", "quote": "parakeet for now, adding models", "excerpt": "fully local transcribe.cpp in your browser at https://mic.sloop.ai/ github repo soon, finalizing ggml / emscripten magic <> simd, pthreads, etc. parakeet for now, adding models, webgpu, wasi server, & electron for native", "context": "", "use_case": "Speech recognition model", "switched_from": "", "story": "Transcribe.cpp"}], "threads": [{"url": "https://news.ycombinator.com/item?id=48894752", "title": "Apple's new SpeechAnalyzer API, benchmarked against Whisper and its predecessor", "mentions": 13}, {"url": "https://news.ycombinator.com/item?id=47666024", "title": "Show HN: Ghost Pepper \u2013 100% local hold-to-talk speech-to-text for macOS", "mentions": 9}, {"url": "https://news.ycombinator.com/item?id=50008427", "title": "Whistle: Speech to Text in 16.9 MB", "mentions": 8}, {"url": "https://news.ycombinator.com/item?id=46731068", "title": "Ask HN: What's the current best local/open speech-to-speech setup?", "mentions": 8}, {"url": "https://news.ycombinator.com/item?id=47143755", "title": "Show HN: Moonshine Open-Weights STT models \u2013 higher accuracy than WhisperLargev3", "mentions": 7}, {"url": "https://news.ycombinator.com/item?id=47040375", "title": "Show HN: Free Alternative to Wispr Flow, Superwhisper, and Monologue", "mentions": 7}], "sentiment_by_month": {"2025-11": {"recommend": 4, "neutral": 2, "warn": 1, "compare": 1}, "2025-12": {"recommend": 3, "neutral": 3, "warn": 1}, "2026-01": {"recommend": 4, "neutral": 2, "compare": 2}, "2026-02": {"recommend": 12, "neutral": 4, "warn": 1}, "2026-03": {"recommend": 8, "warn": 1}, "2026-04": {"recommend": 7, "neutral": 3}, "2026-05": {"neutral": 1, "recommend": 1, "compare": 1}, "2026-06": {"recommend": 5, "warn": 1, "compare": 1, "neutral": 3}, "2026-07": {"recommend": 11, "neutral": 4, "compare": 1, "warn": 2}, "2026-08": {"neutral": 2, "recommend": 2, "warn": 1}, "2026-09": {"recommend": 4, "neutral": 1}, "2026-10": {"recommend": 8, "compare": 1, "neutral": 5}}, "versions": [], "version_unnamed": 73, "family_ref": null, "runs_on": null, "lines": [], "quotes": [{"id": 804596625, "item_id": 50287289, "url": "https://news.ycombinator.com/item?id=50287289", "by": "cjus", "date": "2026-10-29", "version": "2", "stance": "neutral", "quote": "transcribes with NVIDIA's Parakeet TDT 0.6B v2", "excerpt": "I built audio-chat-box to give a program a voice without sending audio anywhere. It reads text aloud with Kokoro, transcribes with NVIDIA's Parakeet TDT 0.6B v2, and can listen hands-free: your program streams microphone samples, Silero VAD finds each thing said, and Parakeet transcribes it when the speaker pauses. It all runs locally through sherpa-onnx. To try it without building anything, there's a small browser d", "context": "", "use_case": "", "switched_from": "", "story": "Show HN: Audio-chat-box \u2013 offline speech in and out for any program, over stdio"}, {"id": 804507330, "item_id": 50281708, "url": "https://news.ycombinator.com/item?id=50281708", "by": "betterproblems", "date": "2026-10-28", "version": "3", "stance": "neutral", "quote": "I used nemotron 3.5 int4 and parakeet 0.6 tdt v3 int4 as the transcriber.", "excerpt": "This is pretty cool. How do I get my own app (app.usequickspeak.fyi) on the website? I used nemotron 3.5 int4 and parakeet 0.6 tdt v3 int4 as the transcriber.", "context": "", "use_case": "transcription", "switched_from": "", "story": "Meeting Notes Takers"}, {"id": 803453922, "item_id": 50215870, "url": "https://news.ycombinator.com/item?id=50215870", "by": "xnx", "date": "2026-10-23", "version": null, "stance": "neutral", "quote": "I rolled my own: podcast-dl + parakeet + Gemini.", "excerpt": "Can't speak for the parent, but they may be using: https://github.com/podly-pure-podcasts/podly_pure_podcasts . I couldn't figure out how to get that working on Windows so I rolled my own: podcast-dl + parakeet + Gemini.", "context": "", "use_case": "", "switched_from": "", "story": "Now Google says it might use your Gmail and Google Photos data to target ads"}, {"id": 803130848, "item_id": 50195678, "url": "https://news.ycombinator.com/item?id=50195678", "by": "marshalla", "date": "2026-10-22", "version": null, "stance": "recommend", "quote": "Parakeet is pretty amazing, though", "excerpt": "\"I've even had it transcribe words before I've finished saying them! \" lol idk if this part is actually a good thing. It's great when you always speak in predictable sandwiches. I do think Parakeet is pretty amazing, though, so some of these product announcements have limited interest to me, especially since it's cloud and costs per hour, even if it's a low cost. Parakeet is trained on text that has had ITN done on i", "context": "", "use_case": "Speech-to-text", "switched_from": "", "story": "Voxtral 3"}, {"id": 803127665, "item_id": 50195479, "url": "https://news.ycombinator.com/item?id=50195479", "by": "marshalla", "date": "2026-10-22", "version": null, "stance": "neutral", "quote": "I've got a custom version of Parakeet Ultra I'm training now", "excerpt": "Neat, can I get mine (yes another one) listed on there? https://mimicscribe.app/ It's just about done. I've got a custom version of Parakeet Ultra I'm training now that does digits better.", "context": "", "use_case": "Speech recognition model", "switched_from": "", "story": "Meeting Notes Takers"}, {"id": 802705664, "item_id": 50169104, "url": "https://news.ycombinator.com/item?id=50169104", "by": "7777777phil", "date": "2026-10-20", "version": null, "stance": "recommend", "quote": "been pretty impressed with the local performance", "excerpt": "I have been using different Parakeet models for STT over the past months and been pretty impressed with the local performance.", "context": "", "use_case": "local speech-to-text", "switched_from": "", "story": "Voxtral 3"}, {"id": 800788304, "item_id": 50049269, "url": "https://news.ycombinator.com/item?id=50049269", "by": "d4rkp4ttern", "date": "2026-10-11", "version": "3", "stance": "recommend", "quote": "dictation was solved a while back with Parakeet V3", "excerpt": "On MacOS, dictation was solved a while back with Parakeet V3, and I haven't seen anything that has a better speed/accuracy tradeoff when talking to AIs, where slight inaccuracies are ok since they can \"read between the lines\". I got annoyed by various UI and other things in tools like VoiceInk, Handy etc and Claude-Coded my own CLI dictation tool called VoxType that uses parakeet-mlx: https://pchalasani.github.io/cla", "context": "", "use_case": "speech recognition", "switched_from": "", "story": "Show HN: BetterWispr \u2013 Free, open-source dictation for Mac"}, {"id": 800378097, "item_id": 50023631, "url": "https://news.ycombinator.com/item?id=50023631", "by": "andai", "date": "2026-10-09", "version": null, "stance": "compare", "quote": "ported my little voice typing gui to use Whistle/Needle instead of Parakeet", "excerpt": "I made a little gui for it. (Or rather, ported my little voice typing gui to use Whistle/Needle instead of Parakeet. Only had to change like 2 lines!) https://gist.github.com/a-n-d-a-i/56ac707eacd8144f83d3651be43b47ef Seems to do roughly as well as parakeet (both speed and quality) so far. I use it on Linux. I think it's cross platform but I haven't tested it. Would love to know! Should be trivial (2 mins of Codex) t", "context": "", "use_case": "speech-to-text", "switched_from": "", "story": "Whistle: Speech to Text in 16.9 MB"}, {"id": 800362640, "item_id": 50022665, "url": "https://news.ycombinator.com/item?id=50022665", "by": "cco", "date": "2026-10-09", "version": null, "stance": "recommend", "quote": "parakeet for STT", "excerpt": "Yeah maybe I'll play around with different precision, I did see improvement when I went from fp4 to full fidelity though. This is my setup for Home Assistant voice, parakeet for STT, GLM 5.3 to generate responses, and Kokoro for TTS. I use local routing for simple commands like turn on the lights, but if I want to ask questions or something, it gets routed to GLM 5.3. It is basically free, half a cent per turn maybe,", "context": "", "use_case": "speech-to-text for Home Assistant voice", "switched_from": "", "story": "Whistle: Speech to Text in 16.9 MB"}, {"id": 800318784, "item_id": 50019924, "url": "https://news.ycombinator.com/item?id=50019924", "by": "gekoxyz", "date": "2026-10-09", "version": null, "stance": "recommend", "quote": "", "excerpt": "The good thing with most NNs is that you can go with fp32 almost always (I put almost because I'm not 100% sure that this covers all the cases) for almost 0 performance degradation. For Parakeet I'm 100% sure because I used it yesterday and it was using just 1GB. Why do you use GLM 5.3 tho? Unless you have an instance running \"free\" 24/7 you could use a local Gemma model. It should be good enough, no? Or maybe with o", "context": "", "use_case": "speech-to-text", "switched_from": "", "story": "Whistle: Speech to Text in 16.9 MB"}, {"id": 800255600, "item_id": 50015975, "url": "https://news.ycombinator.com/item?id=50015975", "by": "aavisangle", "date": "2026-10-09", "version": "3", "stance": "neutral", "quote": "Parakeet TDT 0.6B v3 (int8 CoreML, about 470 MB) runs on-device through FluidAudio's AsrManager", "excerpt": "Parakeet TDT 0.6B v3 (int8 CoreML, about 470 MB) runs on-device through FluidAudio's AsrManager, pinned to CPU plus the Neural Engine.", "context": "", "use_case": "on-device speech recognition", "switched_from": "", "story": "Ask HN: Got a Mac for running agents? what's your setup?"}, {"id": 800242256, "item_id": 50015141, "url": "https://news.ycombinator.com/item?id=50015141", "by": "nl", "date": "2026-10-09", "version": null, "stance": "recommend", "quote": "offloads transcription to Parakeet running in the browser connected to the ESP32", "excerpt": "I recently built a 3D printed ESP32 based transcription device[1] that offloads transcription to Parakeet running in the browser connected to the ESP32. I haven't done my own measurements of error rate but it'd be interesting to see how well this runs on the ESP32 itself. In my case I'm already using one core to run DSP for a beamforming mic array (which works amazingly well for noise cancellation!) so I don't have h", "context": "builds ESP32 transcription device", "use_case": "speech transcription", "switched_from": "", "story": "Whistle: Speech to Text in 16.9 MB"}, {"id": 800221152, "item_id": 50013822, "url": "https://news.ycombinator.com/item?id=50013822", "by": "andai", "date": "2026-10-08", "version": null, "stance": "recommend", "quote": "Parakeet is great I use it everyday", "excerpt": "Technical knowhow? Handy is a gui right? (Also Parakeet is great I use it everyday)", "context": "", "use_case": "speech-to-text dictation", "switched_from": "", "story": "Whistle: Speech to Text in 16.9 MB"}, {"id": 800172209, "item_id": 50010763, "url": "https://news.ycombinator.com/item?id=50010763", "by": "jwr", "date": "2026-10-08", "version": null, "stance": "compare", "quote": "Yes, it is much, much faster and smaller, but accuracy matters a lot", "excerpt": "People keep praising Parakeet, but I've found it to be worse than Whisper Large. Yes, it is much, much faster and smaller, but accuracy matters a lot if you are to use dictation regularly and seriously. I ended up having AI optimize Whisper Large and create a plugin for TypeWhisper, and that's what I use (feeding the results through local Qwen 3.8 running under MTPLX).", "context": "", "use_case": "dictation", "switched_from": "", "story": "Whistle: Speech to Text in 16.9 MB"}, {"id": 800162656, "item_id": 50010166, "url": "https://news.ycombinator.com/item?id=50010166", "by": "theturtletalks", "date": "2026-10-08", "version": null, "stance": "recommend", "quote": "Parakeet is the gold standard.", "excerpt": "Parakeet is the gold standard. With models like moonshine and koroko (TTS model), it\u2019s more about embedding the model in the application itself. If you\u2019re using Parakeet, embedding it in the application is not feasible. I use parakeet with superwhisper, and I\u2019m making another app that has SST and TTS built in, and I want to use my downloaded parakeet model, but it seems there\u2019s so many different implementations from ", "context": "building an app with STT and TTS", "use_case": "speech-to-text", "switched_from": "", "story": "Whistle: Speech to Text in 16.9 MB"}, {"id": 800150513, "item_id": 50009407, "url": "https://news.ycombinator.com/item?id=50009407", "by": "cgbur", "date": "2026-10-08", "version": null, "stance": "recommend", "quote": "I use https://github.com/cjpais/Handy on Parakeet streaming", "excerpt": "For essentially infinite and fast dictation I use https://github.com/cjpais/Handy on Parakeet streaming (cohere is far better, but slower and has a token output limit so you cant ramble for many minutes). And then just do a cleanup pass with a cheap LLM, it will in my experience, do far better than trying to voice control to go edit a sentence or change words. I just weave instructions into my writing. I understand t", "context": "", "use_case": "streaming speech recognition", "switched_from": "", "story": "Whistle: Speech to Text in 16.9 MB"}, {"id": 797773937, "item_id": 49860871, "url": "https://news.ycombinator.com/item?id=49860871", "by": "gunalx", "date": "2026-09-26", "version": null, "stance": "recommend", "quote": "parakeet for speed", "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": "fast speech recognition", "switched_from": "", "story": "Ask HN: What would be the best local ASR model"}, {"id": 797045890, "item_id": 49815368, "url": "https://news.ycombinator.com/item?id=49815368", "by": "tefkah", "date": "2026-09-23", "version": null, "stance": "recommend", "quote": "ie using Parakeet", "excerpt": "After using Superwhisper a lot i wanted something more accurate/faster, ie using Parakeet. I found https://opensuperwhisper.com/. It's pretty ok, long startup time however, and something is overriding the default Opt+` shortcut, so I'm using Opt+Space.", "context": "", "use_case": "", "switched_from": "", "story": "Anyone using a Wispr Flow alternative that is non-cloud?"}, {"id": 796988018, "item_id": 49811751, "url": "https://news.ycombinator.com/item?id=49811751", "by": "zenapollo", "date": "2026-09-23", "version": null, "stance": "neutral", "quote": "vanilla parakeet i use in handy", "excerpt": "Not sure if finetune is the right word here, but their default local model seems to be an in-house fork of parakeet (?, maybe whisper) that works a bit better on tech jargon than vanilla parakeet i use in handy. Though has a few other quirks. Take with a grain of salt as I\u2019m a low volume user, but it\u2019s been enjoyable to test. I kept my wife on handy while i test this one.", "context": "low volume user", "use_case": "local speech model", "switched_from": "", "story": "Anyone using a Wispr Flow alternative that is non-cloud?"}, {"id": 795884403, "item_id": 49742775, "url": "https://news.ycombinator.com/item?id=49742775", "by": "nowittyusername", "date": "2026-09-17", "version": "3", "stance": "recommend", "quote": "asr (parakeet v3), tts (omnivoice) take no more then 400-450 ms total", "excerpt": "I been working and making things with AI agents since Windsurf days, here's what works for me after experimenting and working for a while with these things. As far as the agent goes I find whatever the latest OpenAI model is out worked best for me. This company has burned me the least and I like the models. I tried many but consistency of OpenAI models cant be beat IMO. Though we are post honeymoon phase now I feel l", "context": "", "use_case": "speech recognition for voice agents", "switched_from": "", "story": "Show HN: Share your AI Setup, Learn from others"}, {"id": 792279808, "item_id": 49517488, "url": "https://news.ycombinator.com/item?id=49517488", "by": "dghlsakjg", "date": "2026-09-01", "version": null, "stance": "recommend", "quote": "Parakeet for transcription and senior for diarization.", "excerpt": "Parakeet for transcription and senior for diarization. Check out the fluid-ml library which packages this up for ANE very nicely.", "context": "", "use_case": "speech transcription", "switched_from": "", "story": "Apple caught off guard by AI demand for Mac Mini and Mac Studio"}, {"id": 791637873, "item_id": 49477367, "url": "https://news.ycombinator.com/item?id=49477367", "by": "zackify", "date": "2026-08-28", "version": null, "stance": "recommend", "quote": "Also whisper is pretty outdated vs parakeet", "excerpt": "https://voxtype.io/ https://tryvoiceink.com/ There's so many of these, at this point I've seen 10 clones make the front page each time as if there never existed local only options before. Also whisper is pretty outdated vs parakeet", "context": "", "use_case": "", "switched_from": "", "story": "Show HN: FnScribe \u2013 Open-source, offline dictation for macOS"}, {"id": 791599409, "item_id": 49474963, "url": "https://news.ycombinator.com/item?id=49474963", "by": "jwr", "date": "2026-08-28", "version": null, "stance": "warn", "quote": "I keep hearing that Nvidia Parakeet models are better, but they just don't work as well for me", "excerpt": "It's interesting that the results can be so different depending on the person, the use case and even the microphone used. I use dictation a lot, so I try to stay up to date with the latest models as much as I can. So far, for my needs, nothing could beat Whisper Large v3. I keep hearing that Nvidia Parakeet models are better, but they just don't work as well for me, even though they are unquestionably faster. Things ", "context": "", "use_case": "dictation", "switched_from": "", "story": "Gemini-3.5-Transcribe"}, {"id": 791562720, "item_id": 49472670, "url": "https://news.ycombinator.com/item?id=49472670", "by": "Lucasoato", "date": "2026-08-27", "version": "3", "stance": "recommend", "quote": "I love it but it misses the business specific words when in different language.", "excerpt": "I love it but it misses the business specific words when in different language. Sometimes it stretches them out to fit unrelated words in the language of the rest of the conversation. I miss its speed though.", "context": "", "use_case": "speech transcription", "switched_from": "", "story": "Gemini-3.5-Transcribe"}, {"id": 791562320, "item_id": 49472645, "url": "https://news.ycombinator.com/item?id=49472645", "by": "kridsdale1", "date": "2026-08-27", "version": "3", "stance": "neutral", "quote": "I use Parakeet 3.", "excerpt": "I use Parakeet 3. How does that fare in your testing?", "context": "", "use_case": "", "switched_from": "", "story": "Gemini-3.5-Transcribe"}, {"id": 790877217, "item_id": 49429826, "url": "https://news.ycombinator.com/item?id=49429826", "by": "sourcetms", "date": "2026-08-25", "version": null, "stance": "neutral", "quote": "it's a local speech to text tool using NVIDIA Parakeet", "excerpt": "I built a project called Resonant - it's a local speech to text tool using NVIDIA Parakeet and supports a few other languages. Originally started as an experiment to see what could I, and how far could I go building with LLMs to replace my own software stack. I spent a lot of time on positioning and design. The thing is like 10% model 90% unique, random, obscure features I had to run into walls for days to identify h", "context": "", "use_case": "speech recognition model", "switched_from": "", "story": "Ask HN: Those making $500/month on side projects in 2026 \u2013 Show and tell"}, {"id": 785300225, "item_id": 49081264, "url": "https://news.ycombinator.com/item?id=49081264", "by": "d3k", "date": "2026-07-28", "version": null, "stance": "neutral", "quote": "(using Parakeet)", "excerpt": "Seems nice! I have been using Handy https://handy.computer/ for a while and I am happy with it (using Parakeet), this seems quite similar in terms of features, except that is uses Apple's APIs.", "context": "", "use_case": "speech recognition model for dictation", "switched_from": "", "story": "Show HN: Yap \u2013 OSS on-device voice dictation for macOS with no model to download"}, {"id": 784736673, "item_id": 49046042, "url": "https://news.ycombinator.com/item?id=49046042", "by": "janantos", "date": "2026-07-25", "version": null, "stance": "neutral", "quote": "near realtime spoken word translation/transcription project using selfhosted Gemma4 and Nvidia Parakeet LLM models", "excerpt": "Tbh, I don\u2019t want to be part of codeberg anymore, their community moved away from my believes. And as I have another projects being according them not welcomed there because they are tied to LLM ecosystem. I have prof of concept of AI coding agent entirely written in bash and also near realtime spoken word translation/transcription project using selfhosted Gemma4 and Nvidia Parakeet LLM models. In their latest blog t", "context": "", "use_case": "self-hosted transcription", "switched_from": "", "story": "Brow6el: A full-featured web browser for the terminal using Chromium"}, {"id": 784522241, "item_id": 49032640, "url": "https://news.ycombinator.com/item?id=49032640", "by": "treetalker", "date": "2026-07-24", "version": "3", "stance": "neutral", "quote": "", "excerpt": "I\u2019m a dictaphone fan. Even better than the kind shown in the article are the ones with sliding switches. I have an Olympus model. Mine has both internal memory and SD card capability. And similar to what the author aspires to, I record bits and bobs throughout the day and then use the SD card to transfer them to my Mac. From there, they automatically get added to a MacWhisper watch folder and then automatically trans", "context": "", "use_case": "speech-to-text", "switched_from": "", "story": "Calm technologies that excite me"}, {"id": 783497233, "item_id": 48968577, "url": "https://news.ycombinator.com/item?id=48968577", "by": "d4rkp4ttern", "date": "2026-07-19", "version": "3", "stance": "recommend", "quote": "near instant transcription with Parakeet V3", "excerpt": "I use the Hex[1] app on MacOs for near instant transcription with Parakeet V3. This is how I speak to Code agents at least 80% of the time. The idea is - I tend to be lazy if I need to type lots of details, so speaking lets me get into details that I otherwise wouldn\u2019t and this helps give more useful context to the agent. This often tends to be an unstructured brain dump so I sometimes ask the agent to repeat back wh", "context": "", "use_case": "speech transcription", "switched_from": "", "story": "Speech Recognition and TTS in less than 500kb"}, {"id": 782482145, "item_id": 48905134, "url": "https://news.ycombinator.com/item?id=48905134", "by": "rofrol", "date": "2026-07-14", "version": "3", "stance": "warn", "quote": "I speak Polish and it detects Russian.", "excerpt": "App works great. Parkeeter TDT V3 not so much. I speak Polish and it detects Russian. Also there is no possibility to force language: > unfortuntately parakeet-v3 model doesn;t recieve or output language id https://github.com/NVIDIA-NeMo/Speech/issues/14799#issuecomment-3634425420 Also this: > If you are using Parakeet for English only then you should be using V2. V3 is for several languages and is worse at English o", "context": "", "use_case": "Speech recognition", "switched_from": "", "story": "Apple's new SpeechAnalyzer API, benchmarked against Whisper and its predecessor"}, {"id": 782460130, "item_id": 48903758, "url": "https://news.ycombinator.com/item?id=48903758", "by": "hamza_q_", "date": "2026-07-14", "version": "3", "stance": "neutral", "quote": "using Parakeet V3", "excerpt": "Investigated this and it turned out to be an amusing bug: audio decoding was happening three times instead of just once lol. I've put up a PR to remove the wasteful redundant decoding: https://github.com/FluidInference/FluidAudio/pull/799 With the updated PR code, ran a test comparing transcribing (using Parakeet V3) a 1 hr stereo 44.1 kHz mp3 vs the same audio in 16 kHz mono wav format. The result was about 21.3% sl", "context": "", "use_case": "", "switched_from": "", "story": "Apple's new SpeechAnalyzer API, benchmarked against Whisper and its predecessor"}, {"id": 782406273, "item_id": 48900392, "url": "https://news.ycombinator.com/item?id=48900392", "by": "hamza_q_", "date": "2026-07-13", "version": null, "stance": "recommend", "quote": "sped up Parakeet V2 and V3 to 320x and 282x faster than real time", "excerpt": "Recently contributed a patch to FluidAudio that sped up Parakeet V2 and V3 to 320x and 282x faster than real time, respectively: https://github.com/FluidInference/FluidAudio/pull/507 That means one hour of audio transcribed in 11.25 and 12.75 seconds. The Inscribe post doesn't give a speed factor for SpeechAnalyzer. However, this Argmax blog post reports 70: https://www.argmaxinc.com/blog/apple-and-argmax Based on th", "context": "", "use_case": "Speech-to-text model", "switched_from": "", "story": "Apple's new SpeechAnalyzer API, benchmarked against Whisper and its predecessor"}, {"id": 782398833, "item_id": 48899927, "url": "https://news.ycombinator.com/item?id=48899927", "by": "tapland", "date": "2026-07-13", "version": null, "stance": "recommend", "quote": "Parakeet has great mandarin>CN text", "excerpt": "I use Systran/faster-whisper-medium for real-time subtitling, but you need to get used to the context it's used it and the weirdness it translates into. Parakeet has great mandarin>CN text, but running that + a translation model has been tricky and I never got it fast.", "context": "", "use_case": "Mandarin transcription", "switched_from": "", "story": "Apple's new SpeechAnalyzer API, benchmarked against Whisper and its predecessor"}, {"id": 782392945, "item_id": 48899559, "url": "https://news.ycombinator.com/item?id=48899559", "by": "cootsnuck", "date": "2026-07-13", "version": "3", "stance": "recommend", "quote": "Nvidia's Parakeet v3 is good enough for my own just local transcription most of the time.", "excerpt": "For just desktop, I use Handy (open source): https://handy.computer/ And yea, Nvidia's Parakeet v3 is good enough for my own just local transcription most of the time. When I need local transcription to be more reliable and I don't have the energy to proof read a long ramble, I still often just pop open chatGPT, dictate, cut, paste. But we're pretty much already to the point where local transcription models can repla", "context": "", "use_case": "local transcription", "switched_from": "", "story": "Apple's new SpeechAnalyzer API, benchmarked against Whisper and its predecessor"}, {"id": 782381889, "item_id": 48898868, "url": "https://news.ycombinator.com/item?id=48898868", "by": "satvikpendem", "date": "2026-07-13", "version": null, "stance": "warn", "quote": "", "excerpt": "No, there are better open weight models: https://artificialanalysis.ai/speech-to-text/non-streaming Apparently MOSS-Transcribe-Diarize is quite good too as it released only a few days ago.", "context": "", "use_case": "", "switched_from": "", "story": "Apple's new SpeechAnalyzer API, benchmarked against Whisper and its predecessor"}, {"id": 782379585, "item_id": 48898724, "url": "https://news.ycombinator.com/item?id=48898724", "by": "bel8", "date": "2026-07-13", "version": "3", "stance": "recommend", "quote": "It's  great for me when configured to use Parakeet v3.", "excerpt": "Obligatory mention of Handy which is open-source and cross-platform: https://github.com/cjpais/Handy It's great for me when configured to use Parakeet v3.", "context": "", "use_case": "speech transcription", "switched_from": "", "story": "Apple's new SpeechAnalyzer API, benchmarked against Whisper and its predecessor"}, {"id": 782373745, "item_id": 48898359, "url": "https://news.ycombinator.com/item?id=48898359", "by": "wahnfrieden", "date": "2026-07-13", "version": null, "stance": "compare", "quote": "", "excerpt": "I am running the MLX fork https://huggingface.co/vanch007/mlx-MOSS-Transcribe-Diarize Generally labs don't release MLX or ANE versions and we must rely on finding someone who's converted it Parakeet is not multilingual so not directly comparable Where do you see 16GB? MOSS is smaller than Parakeet at 1.82GB", "context": "", "use_case": "", "switched_from": "", "story": "Apple's new SpeechAnalyzer API, benchmarked against Whisper and its predecessor"}, {"id": 782364176, "item_id": 48897761, "url": "https://news.ycombinator.com/item?id=48897761", "by": "dinfinity", "date": "2026-07-13", "version": "3", "stance": "recommend", "quote": "I use Parakeet V3 via this tool and it is actually quite reliable for me (in English)", "excerpt": "I use Parakeet V3 via this tool and it is actually quite reliable for me (in English): https://github.com/cjpais/Handy", "context": "", "use_case": "speech transcription", "switched_from": "", "story": "Apple's new SpeechAnalyzer API, benchmarked against Whisper and its predecessor"}, {"id": 782354816, "item_id": 48897176, "url": "https://news.ycombinator.com/item?id=48897176", "by": "sherlock-holmes", "date": "2026-07-13", "version": null, "stance": "recommend", "quote": "i am using parakeet with https://handy.computer daily and it's amazing.", "excerpt": "i wish they had benchmarked it against parakeet-unified-en-0.6b and cohere-transcribe-03-2026. i am using parakeet with https://handy.computer daily and it's amazing.", "context": "", "use_case": "Speech-to-text dictation", "switched_from": "", "story": "Apple's new SpeechAnalyzer API, benchmarked against Whisper and its predecessor"}, {"id": 782335713, "item_id": 48895982, "url": "https://news.ycombinator.com/item?id=48895982", "by": "trencedamp", "date": "2026-07-13", "version": null, "stance": "recommend", "quote": "it's perfect with parakeet.", "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": "Whisper", "story": "Apple's new SpeechAnalyzer API, benchmarked against Whisper and its predecessor"}, {"id": 782324128, "item_id": 48895258, "url": "https://news.ycombinator.com/item?id=48895258", "by": "coder543", "date": "2026-07-13", "version": null, "stance": "recommend", "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": 780447008, "item_id": 48777938, "url": "https://news.ycombinator.com/item?id=48777938", "by": "randomblock1", "date": "2026-07-03", "version": "3", "stance": "recommend", "quote": "Parakeet TDT v3 is half the size, way faster, and lower WER", "excerpt": "No, there are quite a few models which are smaller, more accurate, and faster. For example Parakeet TDT v3 is half the size, way faster, and lower WER. There's also Voxstral, which is much larger but also even more accurate. But the ecosystem isn't as mature, so Whisper is still a valid option, even now. For example Parakeet uses Nemotron framework (made by Nvdia), normally you need CUDA, so you need to use an ONNX v", "context": "", "use_case": "", "switched_from": "", "story": "Jamesob's guide to running SOTA LLMs locally"}, {"id": 779296673, "item_id": 48706042, "url": "https://news.ycombinator.com/item?id=48706042", "by": "taffydavid", "date": "2026-06-28", "version": "3", "stance": "recommend", "quote": "you can use parakeet v3 which I find way better than whisper", "excerpt": "I love on-device transcription apps, but this looks like yet another AI generated whisper wrapper. I'm put off by the readme which is clearly LLM generated, with emdash all over the place and bits that smell of user questions turned into talking points. > Why not the Fn / key? It's the obvious one-finger choice, but on macOS the Fn key emits no real keypress \u2014 only a hidden hardware flag \u2014 so And > macOS permissions ", "context": "", "use_case": "speech transcription", "switched_from": "", "story": "Yap \u2013 free offline voice dictation for Mac/Windows/Linux (Wispr Flow alt)"}, {"id": 777394448, "item_id": 48587153, "url": "https://news.ycombinator.com/item?id=48587153", "by": "rob", "date": "2026-06-18", "version": null, "stance": "recommend", "quote": "was completely surprised at just how fast it was for being on-device!", "excerpt": "Yep, here's the v2 and v3: https://huggingface.co/nvidia/parakeet-tdt-0.6b-v2 https://huggingface.co/nvidia/parakeet-tdt-0.6b-v3 It's almost instant on my new M5 Max w/ 36GB of memory, but I used both with Handy on my previous 2019 Intel Mac w/ 16GB memory and was completely surprised at just how fast it was for being on-device! Not instant, but only a couple seconds.", "context": "", "use_case": "on-device speech-to-text", "switched_from": "", "story": "DeepSeek Introduces Vision"}]}