No comments yet.
Quick features:
- Hides partial / broken markdown syntax
- Add your own custom components to LLM output
- Throttling smooths out pauses in the LLM’s streamed output
- Renders output at native frame rate
- Code blocks for every language with Shiki
- Headless: Bring your own styles
The first version is react only for now, but I'm planning to support most frontend frameworks.