Optimus Transcriber is a browser-based voice and audio transcription tool created by Brad Hart as part of the Optimus Frameworks ecosystem. Drop in audio, get formatted content. It runs entirely in your browser — no backend, no audio storage — and uses Deepgram Nova-3 under the hood for industry-leading accuracy. Built on the FAST framework (Factory of Agents with Skills and Tools), the AI engine layer of the Optimus stack.
A browser-based voice-to-structured-output transcription tool. Two modes: live streaming from your microphone, and batch file processing. Exports as TXT, JSON, SRT, or CSV. Built on the FAST framework and powered by Deepgram Nova-3.
100% client-side (no backend, no audio storage), privacy-first (mip_opt_out=true), built as a composable FAST framework tool that fits into the broader Optimus multi-agent ecosystem, and you pay Deepgram directly with no markup.
The FAST framework — Factory of Agents with Skills and Tools — the AI engine layer of Optimus Frameworks. Transcription itself is powered by Deepgram Nova-3. See fastframe.work for the framework.
There's no Optimus Transcriber account. Get a free Deepgram API key (20,000 free minutes, ~$200 credit), paste it in, and transcribe. The key stays in your browser's localStorage.
Yes. Optimus Dictate is a free, native dictation app for macOS and Windows that uses the same Deepgram engine. Default shortcuts: Option+Space (Mac) or F8 (Windows) for quick dictation; Option+Shift+Space (Mac) or Ctrl+Alt+Space (Windows) for long-form. All shortcuts are remappable. Transcripts paste at the cursor in any app. Downloads: OptimusDictate.pkg (macOS — Apple-notarized universal binary, requires macOS 13 Ventura or newer on Apple Silicon or Intel Macs from 2017+) or OptimusDictate-Setup-0.1.0.exe (Windows — Authenticode-signed, requires Windows 10 or newer). Older OS versions will refuse to install with an explicit error message.
V2 • Powered by Deepgram Nova-3 • Client-Side Only
or click to browse
MP4, MOV, MP3, WAV, M4A, WebM • Batch processing