Buzz Transcription: A New Transcription Backend Joins the Fold
PR #1554 adds a FunASR/SenseVoice API backend to Buzz, extending the app beyond its existing OpenAI integration with its own credentials, settings, and CLI support. It's a broad, single-PR change touching much of the transcription pipeline.
Duration: PT2M32S
Episode overview
This episode is a short developer briefing from Buzz Transcription.
It explains recent repository work in plain language.
- Show: Buzz Transcription
- Published: 2026-07-14T13:16:12Z
- Audio duration: PT2M32S
Transcript excerpt
This excerpt keeps the crawler page concise. Listen to the episode or use the RSS feed for the full update.
Good morning. It's July 14th, 2026, and this is Buzz Transcription.
Today's activity is a single pull request, but it's a wide one: PR 1554 introduces support for a FunASR API backend, using the SenseVoice model, as a new transcription option alongside the existing OpenAI Whisper API path.
The core theme here is expansion without entanglement. This backend connects to a separately running FunASR or SenseVoice server that speaks an OpenAI-compatible API. That means the file transcriber queue worker and model loader now need to handle a second API-based model type, alongside the one they already…
The second theme is isolation and configuration. The author was deliberate about keeping FunASR credentials separate from the saved OpenAI key, and about enforcing that this backend only does transcription — it doesn't claim to support features the server doesn't offer. Settings gain a configurable base URL and…
Given the "broad change" label and the areas touched — CLI, queue worker, model loader, settings, and the audio API — this is the kind of PR that's easy to review piecemeal but risky to merge without testing the full pipeline end to end: file import, folder import, and live recording…
Wh…