chatkit-integration

Integrating OpenAI ChatKit with a custom backend requires extending the base ChatKit server via a custom subclass to handle user messages, while the base class manages threads and items to preserve conversations across sessions.

Inject user and page context into ChatKit system prompts by implementing page-context extraction and applying RequestContext scoped operations, feeding the extracted metadata directly into the prompt to guide agent behavior.

Persisting ChatKit conversations across sessions with multi-tenant isolation is achieved through RequestContext scoped operations within the custom backend subclass, ensuring thread and item data remain separated per tenant.

Attach authentication headers and user metadata to frontend ChatKit fetch requests by implementing a custom fetch interceptor, which intercepts outgoing requests and injects the required contextual payloads before they reach the server.

Script loading safeguards are necessary for ChatKit frontend integration to ensure reliable rendering and prevent script execution errors during page-context extraction and initial component mounting.

ChatKit server extensions fail to preserve conversation history when the custom implementation bypasses the base class thread and item management, requiring strict delegation to the base class for persistent chat history operations.