machinefriendly

[machinefriendly @machineFriendly]

I kept having the same stupid thought:
ChatGPT's voice dictation is already better than every paid tool I've tried. So why am I stuck using it only on https://t.co/57pAU60nxx?
So I built a free Chrome extension app Micly to fix it.
demo https://t.co/d7qN01d9RL

[machinefriendly @machineFriendly]

@AmyChantal76432 The accuracy gap is the worst part. If you're ever at a desktop, Micly pipes your dictation through ChatGPT's Whisper — handles French and accents far better than native voice-to-text. Chrome only, free https://t.co/57pAU60nxx login. https://t.co/Q3TVWtfbKm

[machinefriendly @machineFriendly]

@JCavanaugh43836 No idea why the native ones keep regressing. If you're on desktop, Micly routes dictation through ChatGPT's Whisper instead — noticeably more accurate. Chrome only, needs a free https://t.co/57pAU60nxx login. https://t.co/Q3TVWtfbKm

[machinefriendly @machineFriendly]

@thagoodamerican login (free tier works). https://t.co/Q3TVWtfbKm

[machinefriendly @machineFriendly]

@MacpreneurFM For the transcription slot — Micly is a free Chrome ext (not a subscription) that uses ChatGPT's built-in Whisper. ≤20 min files. Worth mentioning alongside the subscription bundle. https://t.co/Q3TVWtfbKm

[machinefriendly @machineFriendly]

@TheUfoJoe For the transcribe step — Micly is a free Chrome ext that pipes audio/video into ChatGPT's Whisper. Up to 20 min per clip, so if you cut to the relevant sections you're under the cap. https://t.co/Q3TVWtfbKm

[machinefriendly @machineFriendly]

@LouisAnslow This workflow works now — Micly (free Chrome ext) does the transcribe step via ChatGPT's Whisper, then Claude or ChatGPT does the grouping. Up to 20 min audio. https://t.co/Q3TVWtfbKm

[machinefriendly @machineFriendly]

@gr0uchy_ The 'just transcribe' part is the bottleneck. Micly is a free Chrome ext that does it via ChatGPT's Whisper — drop the audio, get the text, paste into your draft. Up to 20 min per clip. https://t.co/Q3TVWtfbKm

[machinefriendly @machineFriendly]

@asyafrudin For shorter clips (under 20 min), Micly skips the manual part — drop the audio into ChatGPT via this free Chrome ext and you get the transcript via Whisper. No API key. https://t.co/Q3TVWtfbKm

[machinefriendly @machineFriendly]

@RelentlessNik Add to the list: STOP paying for Otter/Rev → use Micly. Free Chrome ext that pipes audio/video files through ChatGPT's Whisper. Up to 20 min, no API key, no signup. https://t.co/Q3TVWtfbKm

[machinefriendly @machineFriendly]

@heynavtoor You just described Micly's pitch. We wrap that free Whisper into a Chrome ext — drop an audio/video file (≤20 min) into ChatGPT, get the transcript, no API key. Same model Otter charges $17 for. https://t.co/Q3TVWtfbKm

[machinefriendly @machineFriendly]

@tulipdotmd The no-code version for non-CLI folks: Micly — Chrome ext that calls ChatGPT's built-in Whisper. Same model, zero setup, 20 min cap. https://t.co/Q3TVWtfbKm

[machinefriendly @machineFriendly]

@arronnes Or skip Descript's $16/mo — Micly is a free Chrome ext that does the transcribe step via ChatGPT's Whisper, then paste into Claude. Up to 20 min. https://t.co/Q3TVWtfbKm

[machinefriendly @machineFriendly]

@Mrxseek Same workflow here. For the "transcribe" step I use Micly — free Chrome ext that pipes the voice note through ChatGPT's Whisper, then I paste into Claude. No API key, no signup. https://t.co/Q3TVWtfbKm

[machinefriendly @machineFriendly]

@heyblake Step 2 is the friction. Built Micly for this: drop the voice memo, ChatGPT's Whisper transcribes, paste into your AI of choice. Free Chrome ext, no API key. https://t.co/Q3TVWtfbKm

[machinefriendly @machineFriendly]

New in Micly: file → transcript 🎙️
Drop any audio/video into ChatGPT and get the text back via Whisper
— No API key
— No signup
— Files never leave your browser
— Up to 25 MB / 20 min
— mp3, m4a, wav, ogg, opus, webm, flac, mp4, mov
Free Chrome extension: https://t.co/Q3TVWtfbKm

[machinefriendly @machineFriendly]

@StretchGoat_ For next time — Micly is a free Chrome ext that solves the "gpt can't" part: pipes the mp3 through ChatGPT's built-in Whisper. Up to 20 min, no signup. https://t.co/Q3TVWtfbKm

[machinefriendly @machineFriendly]

@Oxymarun_ @WhatsApp You can already do this — export the voice note and drop it into Micly (free Chrome ext, uses ChatGPT's Whisper). Up to 20 min. https://t.co/Q3TVWtfbKm

[machinefriendly @machineFriendly]

@MaeDallas lol relatable — built a free Chrome ext for exactly this: drop the voice memo, get the transcript via ChatGPT's Whisper, no signup. https://t.co/Q3TVWtfbKm

[machinefriendly @machineFriendly]

Full config + 5 failed attempts that got me there:

https://t.co/wFJ9glyJ6x

[machinefriendly @machineFriendly]

You're in yazi, you see the file you want, and you'd like to drop it into Slack or attach it to a ChatGPT message. Default yazi can't do that. It only copies the path as text.

4 lines of config later, `c, s` in yazi → Cmd-V in any app → the file lands.

[machinefriendly @machineFriendly]

The trick: pbcopy copies TEXT. Slack's paste needs a file REFERENCE. Only AppleScript writes one:

```
osascript -e 'set the clipboard to POSIX file "/path"'
```

Wrap in a script, bind to `c, s` in keymap.toml, done.