Inbox / 20260825-161725
streamed 2 parts
0:15
um so can you uh add the the vercel gateway support and— yeah just that for now
Can you add the Vercel gateway support? Just that for now.
VoiceVector is a native dictation app for macOS and Windows. Hold a key, talk, and clean text lands wherever your cursor is — terminal, editor, anywhere.
macOS: signed & notarized — opens first try · Windows: one signed exe — no installer, no admin, nothing to unzip
How it works
No app to switch to and nothing to click. Dictation happens in whatever you're already typing in.
Or double-tap to latch for longer dictations. Esc cancels.
Pause to think — finished phrases transcribe in the background while you talk.
Cleaned text is pasted at your cursor. Your clipboard is put back the way it was.
um so can you uh add the the vercel gateway support and— yeah just that for now
Can you add the Vercel gateway support? Just that for now.
Features
Dictation tools keep growing dashboards. VoiceVector grows in the other direction: fewer moving parts, more of them yours.
ElevenLabs, Cerebras, Vercel AI Gateway, or any OpenAI-compatible server — including one on your own machine. API keys live in your keychain, never in a config file.
A few seconds of silence sends the phrase ahead for transcription. Long dictations finish in the time of the last sentence.
Point transcription and cleanup at the same audio-capable model and VoiceVector sends your voice and the cleanup prompt in a single request.
The exact cleanup instructions are visible and editable in settings. Add your vocabulary — names, jargon, product words — and the models respect it.
Every recording is a .wav, every transcript a
.md in a folder you can see. Grep them, sync them, or
forward a folder to a webhook.
Zero third-party dependencies. Signed and notarized on macOS. One-click updates that verify the publisher before installing. Source available, all of it — free for individuals and organizations under 1,000 people.
Local first
VoiceVector is just a viewer over plain files. Delete the app and every dictation you ever made is still sitting in your documents, readable by anything.
---
date: 2026-08-25T21:17:40Z
duration: 15.1
audio: 20260825-161725.wav
stt: ElevenLabs/scribe_v2
cleanup: Cerebras/gpt-oss-120b
status: complete
---
Can you add the Vercel gateway support? Just that for now.
## Raw transcript
um so can you uh add the the vercel gateway support and— yeah
just that for now