Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Show HN: Whispering – Open-source, local-first dictation you can trust

github.com

Post reply on HN
Print view View original

151–159 of 159 posts

Re: Show HN: Whispering – Open-source, local-first dictation you can trust

#151

Post by SebKba »

wkcheng wrote: post #63
daemonologist wrote: post #43

Earlier quoted context omitted.

Parakeet is amazing - 3000x real-time on an A100 and 5x real-time even on a laptop CPU , while being more accurate than whisper-large-v3 ( https://huggingface.co/spaces/hf-audio/open_asr_leaderboard ). NeMo is a little awkward though; I'm amazed it runs locally on Mac (for MacWhisper).

Yeah, Parakeet runs great locally on my M1 laptop (through MacWhisper). Transcription speed of recordings feel at least 10x faster than Whisper, and the accuracy is better as well. Push to talk for dictation is pretty seamless since the model is so fast. I've observed no downside to Parakeet if you're speaking English.

Parakeet v3 supports many more languages. Works great with MacWhisper.

Re: Show HN: Whispering – Open-source, local-first dictation you can trust

#152

Post by rpdillon »

yunohn wrote: post #39
rpdillon wrote: post #18

Earlier quoted context omitted.

The docs are pretty clear that you need to use speaches if you want entirely local operation. https://speaches.ai/

It’s not very clear, rather just a small mention. Given OP’s extensive diatribe about local-first, the fact that it prefers online providers is quite a big miss tbh.

Yeah, apologies, I did read through the vast majority of the readme before replying. I realize this is not common at all.

Re: Show HN: Whispering – Open-source, local-first dictation you can trust

#153

Post by braden-w »

PickledJesus wrote: post #124

Great software, I've been using this since the start of this year, I use it every day, initially as a frustration with ChatGPT and Claude not having proper voice support in their desktop versions and then everywhere. When you are in an environment where you can dictate, it really is a game changer. Not only is dictating much faster than typing, even if you're a fast typist, I find that you don't have the sticking pro…

Thank you for the support! I'm glad to hear that it's been helping you since the start of the year. Totally agree on the transformation prompts. It's challenging to get the transformation model to not occasionally get short-circuited, especially when I end up having it format a dictated prompt. Instead of formatting, it executes the prompt.

Sorry to hear about the auto-paste feature and taskbar icons. We'll try to restore these in the future, and you can track taskbar here:

https://github.com/epicenter-so/epicenter/issues/607

Re: Show HN: Whispering – Open-source, local-first dictation you can trust

#154

Post by braden-w »

danmeier wrote: post #141

Hot take: I think all these dictation tools are solving the wrong problem: they're optimizing for accurate transcription (and latency) when users actually need intelligent interpretation. For example: People don't speak in perfect emails. They speak in scattered thoughts and intentions that require contextual understanding.

I totally agree with this hot take. Whispering is not there yet, but I eventually want it to store as many of the transcripts as plain text markdown, alongside your audio files, in a folder.

The idea is that as we add more local-first apps into the ecosystem (writing, etc.), they're share this context. Transcription would benefit immensely if you also had a writing app that you could trust to store your data. To execute that vision, we needed a transcription app where we have control over how data is stored, and the best solution was to build our own.

Re: Show HN: Whispering – Open-source, local-first dictation you can trust

#155

Post by fencepost »

braden-w wrote: post #111
braden-w wrote: post #71

Earlier quoted context omitted.

Need to run a diff against 7.2.2 tag against 7.3.0; I suspect the issue might be something related to an edit I made on `tauri.conf.json` or one of my Rust dependencies.

We're actively tracking this issue here: https://github.com/epicenter-so/epicenter/issues/440 Thank you again for bringing this to my attention! Need to step up my Windows development.

Keep in mind that AVG and Avast are owned by the same company now so overlap is likely.

I'm not at my computer but IIRC there was mention of connecting out to a couple of initially unidentifiable domains but a little digging makes it seem they're poorly documented but related to visual studio analytics. I ignored after seeing that.

Re: Show HN: Whispering – Open-source, local-first dictation you can trust

#156

Post by genewitch »

generalizations wrote: post #140
genewitch wrote: post #134

Earlier quoted context omitted.

before i even bother opening that github: does it work on windows? so far, all of the whisper "clones" run poorly, if at all, on windows. I do have a 3060 and a 1070ti i could use just for whisper on linux, but i have this 3090 on my windows desktop that works "fine" for whisper, TTS, SD, LLM. Whisper on windows, the openai-whisper, doesn't have these q8_0 models, it has like 8 models, and i always get an error about…

If you want to stay near the bleeding edge with this stuff, you probably want to be on some kind of linux (or lacking that, Mac). Windows is where stuff just trickles down to eventually.

you're not wrong; i really should try just running linux and seeing how good the steam gaming layer is these days. And if SDR# runs on linux under wine or whatever.

Re: Show HN: Whispering – Open-source, local-first dictation you can trust

#157

Post by ajolly »

The killer feature I'm still looking for is software that will do voice to text, but insert the text into the text box that was active when I started talking, not ended talking.

That way I could click on a text box start talking but have the rest of my brain switch to doing a second task.

Re: Show HN: Whispering – Open-source, local-first dictation you can trust

#158

Post by emacsen »

braden-w wrote: post #110
emacsen wrote: post #62

Earlier quoted context omitted.

Same with the deb. :(

Thanks for flagging this, and sorry that this is happening! Does downloading the model manually work? I wonder if it's related to this: https://github.com/epicenter-so/epicenter/issues/669

I don't think it's the same error, but without a good error message I don't know.

I did manually download the models and associated them, which are great but then the audio didn't work. On the browser version, it never asks me for permission for an audio device, and on the native version, it makes a file of 0 length and then complains it can't read the contents.

My read is that the project looks very interesting, and I'd love a FLOSS replacement for Aqua Voice, but this software isn't ready for everyday use yet, at least not on Linux.

I'd love to help somehow, whether that's a donation or experimenting if you point me to somewhere.

Re: Show HN: Whispering – Open-source, local-first dictation you can trust

#159

Post by Aachen »

0xbadcafebee wrote: post #81
Aachen wrote: post #59

Earlier quoted context omitted.

Will give it a spin, thanks!

I also just found something that sounds genuinely realistic: Piper ( https://github.com/OHF-Voice/piper1-gpl/tree/main ). It's slow but apparently you can run it as a daemon to be faster, and it integrates with Home Assistant and Speech Dispatcher. $ sudo apt update $ sudo apt install -y python3 python3-pip libsndfile1 ffmpeg $ python -m venv piper-tts $ ./venv/piper-tts/bin/pip install piper-tts $ ./venv/piper-tts/b…

I've tried Piper using this app: https://f-droid.org/packages/org.woheller69.ttsengine

It has some mispronounciations in the texts I tried to listen to, besides using so much RAM that it kills basically all other apps to make space for this. Not really worth it when espeak is already understandable

Festival I've tried in the meantime and doesn't support Dutch or German, two of the three languages I use regularly. I keep coming back to espeak at the only option that will simply always work xD

Post reply on HN