Stage 2 · v1.5In development

Auto-subtitles that never leave your machine.

On-device transcription and captioning — model choice, voice activity detection, styling presets, and a correction pass before anything touches your timeline. No cloud calls.

Correction pass
00:04.2Welcome back — today we're grading the opening scene.
00:08.9Every word of this was transcribed on this machine.
00:13.1No audio ever leaves your computer.

Feature roadmap — everything planned, honestly labeled

What Hermite Subs will do

On-device transcription

Choice of model size and family — small and fast, or larger and more accurate — running entirely on your machine.

In development

Voice activity detection

Silence is skipped before transcription, so long timelines process faster.

In development

Styled subtitle tracks

Auto-generated Resolve subtitle/Text+ tracks with adjustable styling presets — font, position, background, karaoke-style highlight.

Coming soon

Correction pass

A waveform + text editor to fix errors before committing to the timeline. Nothing is written without your review.

Coming soon

Speaker diarization

Identify and label different speakers in the transcript.

Coming soon
On-device means on-device. Transcription runs on your hardware with models you choose and can inspect on disk. There is no upload step, no processing server, no "anonymized sample" — the audio path never touches a network. That's not a settings toggle; it's the architecture.

Watch it come together

Hermite is built in the open — the roadmap and changelog update as the code does.