Chat Endpoint

Public endpoint for AI-powered portfolio chat. No authentication required.

Method	Endpoint	Description
POST	/api/v1/chat	Send chat message

POST /chat

Send a chat message and receive an AI response. The chat service includes input/output guardrails for content safety and PII protection.

Request Body

{
  "message": "Tell me about your TypeScript experience",
  "visitorId": "visitor-unique-id"
}

Field	Type	Required	Description
message	string	Yes	User message (1-2000 chars)
visitorId	string	Yes	Client-generated visitor identifier (1-100 chars)

Query Parameters

Parameter	Type	Default	Description
includeToolCalls	boolean	false	Include tool call details in response

Response

{
  "sessionId": "sess_abc123",
  "message": {
    "id": "msg_xyz789",
    "role": "assistant",
    "content": "I have extensive experience with TypeScript...",
    "createdAt": "2025-01-25T10:00:00Z"
  },
  "tokensUsed": 150,
  "toolCalls": []
}

TIP

The toolCalls field is only included when includeToolCalls=true is passed as a query parameter.

Error Responses

• 429 Too Many Requests - Rate limited
• 502 Bad Gateway - LLM service unavailable

See Also

• MCP Server & AI Tools - Tools available to the chat AI
• Conventions - Rate limiting details
• Schemas & Errors - ChatResponse schema and error codes