chatkit-server-implementation

Streaming AI responses use server-sent events (SSE) to push live agent output to clients. This Skill provides a blueprint for implementing SSE event formatting, integrating OpenAI's Agents SDK, and wiring a ChatKit frontend to receive and display streamed messages in real time.

Thread persistence stores conversation history durably so users can resume chats. This Skill covers designing thread storage, retrieving past messages, and passing thread identifiers between frontend and backend to maintain continuous context across reconnections.

Yes. This Skill shows how to flow frontend context—page location, selected text, user metadata—explicitly into agent tool parameters so tools receive precise, actionable input without requiring users to repeat context manually.

RAG retrieves relevant documents or data before generating responses, improving accuracy. This Skill demonstrates plugging a retrieval backend into agent workflows so the agent can search and cite sources while answering questions.

Embedding ChatKit requires a backend that handles SSE streaming, manages threads, and routes context to agents. This Skill provides the full server blueprint—request handling, event formatting, and integration patterns—for production-ready deployment.

Thread storage requires a persistent data layer; the specific choice depends on your stack. This Skill focuses on the thread persistence contract—what data to store and how to retrieve it—leaving database selection to your infrastructure.