Agora’s skip_patterns keep code out of TTS playback
Agora has published a guide on using `skip_patterns` within its Conversational AI platform. This feature allows users to prevent the AI's Text-to-Speech (TTS) from vocalizing structured data like code. The `skip_patterns` are used to manage how the AI processes and presents information during UI previews, tagging, and profile updates.
Key Takeaways
- `skip_patterns` in Agora Conversational AI stop TTS from vocalizing structured data such as code.
- Agora says the same feature is used for UI previews, tags, and profile updates.
- The guidance is framed as a way to separate speech output from structured data handling.
Why It Matters
Agora is documenting a practical control for Conversational AI users who need TTS to skip code and other structured text instead of reading it verbatim. That matters for voice interfaces where raw markup or code can make playback unusable. The article also shows Agora positioning `skip_patterns` beyond speech, including UI previews, tagging, and profile updates, which suggests the same parsing layer feeds multiple product surfaces. The next concrete signal to watch is how Agora describes additional `skip_patterns` use cases in the product docs or examples.
Read full article at prod.agora.io