[cdliao]

Docs for ZoomAIRealtimeLLMService (pipecat-ai/pipecat#6353), the speech-to-speech service for Zoom's AI Voice Agent API.

Changes

• New page api-reference/server/services/s2s/zoom-ai.mdx, in STYLE.md's service-page order:
  • Overview, Installation, Prerequisites.
  • Configuration: constructor, Settings, and the Provider, Listen and Speak types.
  • Usage: basic setup, voice and vocabulary, your own OpenAI or Anthropic model, updating settings mid-conversation.
  • Notes and Event Handlers.
• Added to the Realtime LLM navigation group and the Supported Services table.
• llms.txt and llms-full.txt regenerated.

Checks

• node scripts/docs-meta-lint.mjs: 0 errors.
• check-reference.py and check-imports.mjs against the code PR: no findings for the new page.
• Prettier applied.
• Every code sample was run against the production Voice Agent.