Every word transcribed.
Not one word uploaded.
my_words_ turns your recordings, meetings and media into transcripts and summaries on your own machine. Local Whisper does the listening, a local AI model does the summarising, and the files land on your disk. No cloud, no account, no per-minute meter. Free.
Your voice is training data to them.
Cloud transcription is a simple trade: you hand over the audio, they hand back the text, and your recordings sit on their servers under their terms. The client call, the standup, the interview, the note you mumbled to yourself in the car. my_words_ refuses the trade: the audio never leaves the machine that recorded it.
Local Whisper transcription
Whisper-family models run on your own hardware. The same model family the cloud services build on, minus the upload.
Local AI summaries
Long transcript in, tight summary out, produced by a local model on your machine. No cloud model reads your meetings.
No account, no meter
No sign-up, no per-minute billing, no monthly cap ticking down while you talk. It is your computer doing the work.
Files, not a feed
Transcripts and summaries land as plain files on your disk. Grep them, archive them, feed them to whatever you like. They are yours.
Point it at a file, a folder, or a URL you have rights to.
Drop in a recording, or pull media in from a link. For sites you are logged in to, my_words_ can use your own browser session, so your access stays your access, on your machine.*
Your recordings
Meetings, voice notes, interviews, lectures you recorded. Point my_words_ at the file and get text back in minutes.
Media by link
Give it a URL to media you own or have the right to use, and it fetches, transcribes and summarises in one pass.
Your own browser session
For content behind your own login, it borrows the session from your browser. No credentials handed to a third party, because there is no third party.
Pick your depth
Choose the Whisper model size for the speed-accuracy trade you want. Bigger models, better ears, still local.
* my_words_ is built for your own recordings and media you have the rights to use. You are responsible for respecting the source platform's terms and copyright law for anything you point it at.
my_words_ vs cloud transcription, on the facts
The cloud services are polished. The difference is where your audio goes, who keeps it, and what you pay for the privilege.
| my_words_ (Local Gorilla) | Otter.ai | Rev | |
|---|---|---|---|
| Where the audio goes | Nowhere. Stays on your machine | Uploaded to their cloud | Uploaded to their cloud |
| Account required | No | Yes | Yes |
| Pricing | Free | Free tier + subscription | Per-minute / subscription |
| Who keeps the recording | You, as files on disk | Stored on their servers | Stored on their servers |
| Summaries | Local AI model, on-device | Cloud AI on their backend | Cloud AI on their backend |
| Works offline | Yes, for local files | No | No |
| Platforms | macOS (Apple Silicon) | Web / mobile | Web / mobile |
Otter.ai (otter.ai) and Rev (rev.com) are cloud transcription services by category: audio is processed and stored on their servers behind an account, with free tiers rationed by minutes and paid plans beyond. Facts stated at category level; corrections welcome: admin@lionsedge.tech
Free. Actually free.
No subscription, no credit meter, no free tier that runs out mid-meeting. Your machine does the compute, so there is nothing to bill you for. my_words_ is part of the Local Gorilla suite: it shows you what on-device feels like, and it stays free.
- Local Whisper transcription, all model sizes
- Local AI summaries
- Files, folders, and URLs you have rights to
- No account, no upload, no telemetry
Questions
Is there a local, private alternative to Otter or Rev?
Yes. my_words_ transcribes your recordings on your own machine using local Whisper models, then summarises them with a local AI model. Nothing is uploaded, there is no account, and it is free. Cloud transcription services process and store your audio on their servers behind an account and a subscription.
Does my audio leave my machine?
No. Transcription runs on local Whisper models and summaries run on a local AI model, both on your own hardware. There is no my_words_ server, no upload step and no telemetry. The transcripts and summaries land as files on your machine and stay there.
How much does my_words_ cost?
my_words_ is free. There is no subscription, no credit meter, no per-minute billing and no account. Cloud transcription services typically meter you by the minute or ration a free tier behind a monthly plan.
Can I transcribe anything I find online?
my_words_ is built for your own recordings, your meetings, and media you have the rights to use. When you pull media in from a URL, you are responsible for having the right to use that content and for respecting the source platform's terms and copyright law. The tool runs locally; what you feed it is your call and your responsibility.
How accurate is local transcription?
my_words_ runs Whisper-family models locally, the same model family many cloud services build on. Accuracy depends on the model size you pick and your audio quality; on Apple Silicon the larger models run comfortably and handle accents, including Australian English, well.
What platforms does it run on?
macOS on Apple Silicon first. Because everything runs locally, there is no service dependency: your machine is the whole stack.
Your words. Your machine. Full stop.
Free, on-device transcription and summaries. Get told the moment it lands.
Notify me