Skip to main content

Chat Playground

What this page is for​

Query your published knowledge base using RAG-powered chat with citations and conversation history.

Endpoint​

  • POST /api/collections/:collectionId/chat

How it works​

  1. Your message is embedded and used to search the collection's published chunks.
  2. Relevant chunks are retrieved and passed as context to the LLM.
  3. The LLM generates an answer grounded in your knowledge base.
  4. Citations link back to the source chunks with relevance scores.

Steps​

  1. Ensure the target collection has published chunks.
  2. Send a chat message with an optional sessionId for conversation continuity.
  3. Review the answer and its citations.
  4. Continue the conversation using the returned sessionId.

Example​

First message (new conversation)​

curl -X POST http://localhost:3000/api/collections/<collectionId>/chat \
  -H "Content-Type: application/json" \
  -H "X-User-ID: user-1" \
  -d '{
    "message": "How does authentication work in our system?"
  }'

Response​

{
  "answer": "Authentication uses JWT tokens issued by...",
  "sessionId": "chat-session-456",
  "citations": [
    {
      "chunkId": "chunk-789",
      "content": "The authentication module issues JWT tokens...",
      "score": 0.92
    }
  ]
}

Follow-up message (same conversation)​

curl -X POST http://localhost:3000/api/collections/<collectionId>/chat \
  -H "Content-Type: application/json" \
  -H "X-User-ID: user-1" \
  -d '{
    "message": "What about refresh tokens?",
    "sessionId": "chat-session-456"
  }'

Request body​

FieldTypeRequiredDescription
messagestring (1–5,000)yesThe question or prompt
sessionIdstringnoExisting session ID for conversation continuity

Verify​

  • Response contains answer, sessionId, and citations array.
  • Citations reference actual chunk IDs from the collection.
  • Passing the same sessionId maintains conversation context.

Next steps​