The problem
Turning reports and articles into audio content took hours of scripting and recording.
What we built
- File upload and source selection, so each episode uses exactly the material chosen.
- Parameters for length, tone and number of speakers.
- A pipeline that writes a dialogue script with an LLM and converts it to audio with text-to-speech.
Result
A workflow that produces a first-draft episode in minutes, which editors then review and publish.