AI Podcast
AI Podcast turns materials you have added, or a topic, into one episode you can listen to. Open it from Apps, or pin it to the sidebar. After you open a material, the learning panel on the right also has an entry.

Make one episode
The page steps are: Content → Script → Voice → Synthesize. Generate the script first before later steps unlock. While generating, you can see unlocked steps; editing is temporarily disabled.
- Prepare content. You can enter only a topic, or cite materials or notes. Multiple materials are allowed; only one topic. If you have no materials yet, import them in My Materials first.
- Choose a scene and length, confirm the prompt, then generate the script.

- Review the script. Make edits here before you choose voices.
- Choose voices, then synthesize. Two-person scenes need two voices; one-person scenes need one.

Episodes appear under My Podcasts. Statuses include Draft, Synthesizing, Ready, and Failed. Unfinished drafts can be edited later.
You can also play the current podcast full screen, like a music player:

Scenes and length
Scenes follow the options on the page. Current choices include:
- Two-person dialogue, solo deep dive, news brief, Q&A flashcards, in-depth interview
- Roundtable, debate of views, storytelling, highlights overview, how-to tutorial
Length has three tiers, as labeled on the page: quick highlights about 3–5 minutes, deep dive about 8–15 minutes, and debate about 5–10 minutes.
Setup
Script and voiceover are configured separately under Settings → AI & models. Podcast generation currently supports ListenHub, Tencent Cloud, and Volcano Engine. Text-to-speech currently supports MiniMax, Bailian, Volcano, and iFlytek. The page has two entries: Podcast generation and Voiceover. Until setup is complete, the page shows Not configured, and that step will not finish on its own.
Learn with the podcast
After the same content is synthesized, you can still turn it into knowledge cards, a quiz, notes, or a presentation.

Related: Built-in apps · Podcast generation key · Text-to-speech key
