Skip to main content

​
Introduction

The Bedrock Chat API provides programmatic access to conversational AI powered by Amazon Bedrock. This RESTful API allows you to create conversations, manage bots, and integrate chat functionality into your applications.

​
Base URL

All API requests should be made to your deployed API Gateway endpoint: 
https://api.example.com/

​
Authentication

The Bedrock Chat API uses two authentication methods depending on the endpoint:

​
1. Cognito JWT Authentication

For regular application endpoints, authentication is done using AWS Cognito JWT tokens. Required Headers: 
Authorization: Bearer <jwt_token>
Token Verification: Tokens are verified against AWS Cognito using the JWKS endpoint: 
https://cognito-idp.{region}.amazonaws.com/{user_pool_id}/.well-known/jwks.json
Token Requirements:
  • Algorithm: RS256
  • Audience: Must match the Cognito Client ID
  • Valid signature from Cognito

​
2. API Key Authentication

For published bot APIs, authentication uses API keys created through the bot publication flow. Required Headers: 
x-api-key: <api_key>
API keys can be created and managed through the publication endpoints. See Published API Endpoints for details.

​
Request Format

All POST and PATCH requests must include: 
Content-Type: application/json

​
Response Format

All responses are returned in JSON format: 
{
  "field1": "value1",
  "field2": "value2"
}

​
Error Handling

The API uses standard HTTP status codes:
Status CodeDescription
200Success
400Bad Request - Invalid parameters
401Unauthorized - Invalid or missing authentication
403Forbidden - Insufficient permissions
404Not Found - Resource doesn’t exist
409Conflict - Resource state conflict
500Internal Server Error
Error Response Format: 
{
  "detail": "Error message describing what went wrong"
}

​
Rate Limiting

For published APIs, rate limiting can be configured per bot with:
  • Quota: Request limits per time period (DAY/WEEK/MONTH)
  • Throttle: Rate limit (requests per second) and burst limit
See Bot Store Endpoints for configuring rate limits.

​
Health Check

You can verify API availability using the health check endpoint: 
GET /health
Response: 
{
  "status": "ok"
}

​
Available Models

The API supports the following models:

​
Claude Models

  • claude-v4-opus
  • claude-v4.1-opus
  • claude-v4.5-opus
  • claude-v4-sonnet
  • claude-v4.5-sonnet
  • claude-v4.5-haiku
  • claude-v3.5-sonnet
  • claude-v3.5-sonnet-v2
  • claude-v3.7-sonnet
  • claude-v3.5-haiku
  • claude-v3-haiku
  • claude-v3-opus

​
Mistral Models

  • mistral-7b-instruct
  • mixtral-8x7b-instruct
  • mistral-large
  • mistral-large-2

​
Amazon Nova Models

  • amazon-nova-pro
  • amazon-nova-lite
  • amazon-nova-micro

​
Other Models

  • deepseek-r1
  • llama3-3-70b-instruct
  • llama3-2-1b-instruct
  • llama3-2-3b-instruct
  • llama3-2-11b-instruct
  • llama3-2-90b-instruct
  • gpt-oss-20b
  • gpt-oss-120b

​
Content Types

The API supports multiple content types in messages:
  • text: Plain text content
  • image: Base64-encoded images (Claude 3+ only)
  • attachment: File attachments
  • toolUse: Tool invocation requests
  • toolResult: Tool execution results
  • reasoning: Reasoning traces (for reasoning-enabled models)

​
Next Steps

Conversations

Learn about conversation management endpoints

Bots

Explore bot creation and management

Bot Store

Discover bot store and search functionality

Published APIs

Publish and share bot APIs

Build docs developers (and LLMs) love

Get started for freeTalk to us