004 Teaches Abad4d20

Agent API Intro

This guide teaches LLMs how to interact with IntentKit agent APIs when building applications. The agent API provides endpoints to create chat threads, send messages, and retrieve conversation history.

Base URL and Authentication

All API endpoints are prefixed with /v1/ and require authentication using a Bearer token.

Base URL: http://localhost:8000/v1/ when local development. For production, you will get it when you get the key.

Authentication: Include the agent token in the Authorization header:

Authorization: Bearer <your_agent_token>

How to Get the API Key

IntentKit provides system skills for managing agent API keys. Active the skills in agent, and let it give you the API Keys. You can get two types of API keys with different access levels:

• Private API Key (sk-): Can access all skills (public and owner-only)
• Public API Key (pk-): Can only access public skills

Quick Start Example

Here's a complete example showing how to create a thread, send a message, and list messages:

1. Create a Chat Thread

curl -X POST "http://localhost:8000/v1/chats" \
  -H "Authorization: Bearer <your_agent_token>" \
  -H "Content-Type: application/json"

Response:

{
  "id": "chat_123456",
  "agent_id": "agent_789",
  "user_id": "agent_789_user123",
  "summary": "",
  "rounds": 0,
  "created_at": "2024-01-01T12:00:00Z",
  "updated_at": "2024-01-01T12:00:00Z"
}

2. Send a Message

curl -X POST "http://localhost:8000/v1/chats/chat_123456/messages" \
  -H "Authorization: Bearer <your_agent_token>" \
  -H "Content-Type: application/json" \
  -d '{
    "message": "Hello, how can you help me today?"
  }'

Response:

[
  {
    "id": "msg_user_001",
    "agent_id": "agent_789",
    "chat_id": "chat_123456",
    "user_id": "agent_789_user123",
    "author_id": "agent_789_user123",
    "author_type": "API",
    "message": "Hello, how can you help me today?",
    "created_at": "2024-01-01T12:00:01Z"
  }
]

Advanced Features

About User ID

All user_id parameter in the Agent API is optional. If not provided, the API will use owner role. When you include the user_id in the request, the API will use the public role, and the user can have it's own chat threads. You can use the user_id of your own system to identify the user. If the app is anonymous, you can use javascript library fingerprintjs to generate the user_id.

Streaming Responses

Set stream: true in the message request to receive streaming responses, your client need to support HTTP2 server push.

Attachments

Include images, or files in messages:

{
  "message": "Convert the image to Cyberpunk style",
  "attachments": [
    {
      "type": "image",
      "url": "https://your.image.url"
    }
  ]
}

Search and Super Mode

Search mode will active the model native search, only supported by the model with search ability, like gpt and grok. Super mode will give the agent more step limit, to handle complex task.

Error Handling

The API returns standard HTTP status codes:

• 200: Success
• 201: Created
• 204: No Content
• 400: Bad Request
• 401: Unauthorized
• 404: Not Found
• 500: Internal Server Error

Example error response:

{
  "detail": "Chat not found"
}

OpenAI Compatible API

IntentKit also provides an OpenAI-compatible API endpoint that allows you to use your agent with any OpenAI-compatible client or SDK. This is particularly useful for integrating with existing applications that already use the OpenAI API format.

For detailed, see the OpenAI Compatible API Documentation.

API Reference

For complete API documentation with interactive examples, visit: http://localhost:8000/v1/redoc

If you are using LLM to generate code, providing it with this raw OpenAPI specification link is sufficient:

http://localhost:8000/v1/openapi.json