Skip to content

API Documentation

The generative AI agent exposes a FastAPI-based RESTful API for interacting with the model. This document provides an overview of the API, key endpoints, and how to access the full API documentation.

Overview

The API allows you to interact with the AI agent via HTTP requests. It supports:

  • Sending messages to the AI agent.
  • Streaming conversational responses.
  • Providing contextual data for more tailored responses.

Base URL

The API server runs on the following base URL when deployed locally: 

http://127.0.0.1:8000

For production deployments, replace the base URL with your server's address.

Full API Documentation

The full API documentation is available in OpenAPI format. When the app is running, you can view it interactively using:

To access the OpenAPI JSON specification directly: - Download openapi.json

Key Endpoints

1. POST /v1/chat/completions

Streams a conversational response from the AI agent.

Request

  • Headers:
  • Content-Type: application/json
  • Body: 
    {
  "messages": [
    {
      "role": "user",
      "content": [
        { "text": "Hello, AI!" }
      ]
    }
  ],
  "context": {
    "values": [
      { "key": "session_id", "value": "abc123" }
    ]
  },
  "include_status_updates": false,
  "stream": true
}

Response

  • Success: 
    {
  "id": "abc123",
  "object": "thread.message.delta",
  "thread_id": "None",
  "choices": [
    {
      "delta": {
        "role": "assistant",
        "content": "Hello! How can I assist you?"
      }
    }
  ]
}
  • Error: 
    {
  "error": "Invalid input format",
  "choices": [
    {
      "delta": {
        "role": "assistant",
        "content": "There seems to be an issue with the provided request format. Please check your input and try again."
      }
    }
  ]
}

2. GET /docs

Provides interactive Swagger documentation for exploring the API.

3. GET /redoc

Provides a ReDoc-based view of the API schema.

Using the OpenAPI Schema

The OpenAPI JSON file can be integrated with tools like:

  • Postman: Import the schema to explore and test endpoints.
  • Swagger Editor: Visualize and validate the schema.

For instructions, see Swagger Editor or Postman.

Troubleshooting

  • If you encounter issues accessing the API or endpoints:
  • Ensure the server is running (uvicorn main:app or via Docker/Podman).
  • Verify the base URL and port are correct.
  • Check logs for errors in the server console.

For further details, consult the Getting Started Guide or the Configuration Guide.