Based on current runtime
Two live stages with a manual review between them
ScribeBot accepts voice messages and audio files and returns a transcript. PostBot accepts reviewed text through /repurpose and creates four platform drafts. No current contract transfers the voice note or transcript between the bots automatically.
This is a practical workflow, not a transcription-accuracy benchmark.
Practical workflow
- 1. Transcribe: Send a voice message, audio file, or video note to @vustScribeBot. It returns a transcript, summary, and action items.
- 2. Clean the source: Remove repetition and filler, then check names, numbers, links, and claims. Spoken confidence is not evidence.
- 3. Create variants: Separately send /repurpose plus the reviewed text to @vustPostBot. It returns Telegram, LinkedIn, X, and Instagram drafts.
- 4. Review and publish: Choose a version, restore any missing context, and publish it yourself. /repurpose does not auto-post to external networks.
- Based on
- ScribeBot and /repurpose runtime contracts, with no accuracy claim.
- Best for
- Ideas captured on the move, founder notes, and spoken channel outlines.
- Not ideal for
- Verbatim transcripts or unattended publishing.
- In short
- Transcribe, review, adapt, then publish.
P1 · planned, not live
Direct voice input should remove the manual handoff
The planned PostBot extension is to accept a voice note, reuse Scribe's bounded STT path, show the transcript for confirmation, and only then enter the existing /repurpose pipeline. Until that ships, do not send a voice note to PostBot expecting ready drafts.
This does not require a separate bot.