Every episode starts somewhere.
Blog2Podcast uses source articles from engineering blogs and text submitted by approved members. A source link and publisher label accompany public articles so you can read the original in context.
How material becomes audio
- Capture the source text and its original link.
- Create a summary and two-speaker script using Ollama Cloud.
- Check the script structure and source quotations, then render synthetic speech with Kokoro on the application server.
- Preview the private draft. An administrator reviews it before public release.
Automated checks do not prove every claim is correct. Names, pronunciation, numbers and important caveats need human review. The original article remains the reference.
Permission and attribution
Educational purpose does not grant permission to reuse someone else’s writing. Members must have permission to process submitted text. Source discovery through a feed or website does not establish narration or republication rights. Public release requires a separate rights review.
Raw article bodies are not included in the public API. Follow the original link for the complete article. Publisher names identify the source; they are not partners or endorsements.
Speech credits
New draft speech uses the Kokoro model and Kokoro ONNX runtime integration with stock synthetic voices. Existing recordings may use the earlier speech pipeline.