Uoink your podcasts too.
Add an RSS feed, transcribe episodes locally, and file speaker-aware transcripts beside your video corpus.
RSSWhisper# Episode title feed, episode URL, duration, publish date ## Speakers Speaker 1 / Speaker 2 / Speaker 3 ## Transcript [14:22] Speaker 2: the claim is... ## Entities companies, tools, people, topics
Same source card, longer runtime.
Podcasts sit in the same local library as videos. Your agent can search a founder interview, a lecture, and a YouTube teardown with the same tool call.
- 01
Feed in.
Add a public RSS feed or episode URL from the settings page.
- 02
Whisper runs local.
The helper downloads audio and runs transcription on your machine.
- 03
Corpus lands.
Episode markdown, speaker labels, metadata, and entities land in the library.
Private transcription costs time.
Expect compute time for long audio, especially on CPU. Uoink keeps the job in the background and saves the transcript on disk when it finishes.
Useful caveats, kept visible.
Does Uoink identify speakers by real name?
Optional diarization separates speakers as Speaker 1, Speaker 2, and so on. It needs your own Hugging Face token for the speaker model. You can rename speakers later when you know who is speaking.
Does local Whisper need a GPU?
CPU transcription works. Larger models take longer. Dedicated GPUs can speed it up when available.
Are podcast feeds uploaded to Uoink?
Feed polling, audio download, transcription, and corpus writing run from your machine.