Skip to main content

Overview

Conversation endpoints allow you to create chat conversations, send messages, manage conversation history, and retrieve related documents.

Send Message

Request Body

conversation_id
string
required
Unique conversation identifier
message
object
required
Message object containing the user’s input
bot_id
string
ID of the bot to use for this conversation
continue_generate
boolean
default:"false"
Whether to continue generation from the last message
enable_reasoning
boolean
default:"false"
Enable reasoning mode for supported models

Response

conversation_id
string
The conversation identifier
message
object
The bot’s response message
bot_id
string
The bot ID used
create_time
float
Unix timestamp of conversation creation

Example Request

{
  "conversation_id": "01HQ4V8JFBK9PZXQ8YJKMND8GA",
  "message": {
    "role": "user",
    "content": [
      {
        "content_type": "text",
        "body": "What is Amazon Bedrock?"
      }
    ],
    "model": "claude-v3.5-sonnet",
    "parent_message_id": null
  },
  "bot_id": "my-assistant-bot",
  "continue_generate": false,
  "enable_reasoning": false
}

Example Response

{
  "conversation_id": "01HQ4V8JFBK9PZXQ8YJKMND8GA",
  "message": {
    "role": "bot",
    "content": [
      {
        "content_type": "text",
        "body": "Amazon Bedrock is a fully managed service that makes foundation models from leading AI companies available through an API..."
      }
    ],
    "model": "claude-v3.5-sonnet",
    "children": [],
    "feedback": null,
    "used_chunks": null,
    "parent": "01HQ4V8JFBK9PZXQ8YJKMND8GB",
    "thinking_log": null
  },
  "bot_id": "my-assistant-bot",
  "create_time": 1704067200.0
}

Get Conversation

Path Parameters

conversation_id
string
required
The conversation ID to retrieve

Response

id
string
Conversation identifier
title
string
Conversation title
create_time
float
Unix timestamp of creation
message_map
object
Dictionary mapping message IDs to message objects
last_message_id
string
ID of the most recent message
bot_id
string
Bot ID used in this conversation
should_continue
boolean
Whether generation should continue

List All Conversations

Response

Returns an array of conversation metadata objects:
id
string
Conversation identifier
title
string
Conversation title
create_time
float
Unix timestamp of creation
model
string
Model used in the conversation
bot_id
string
Bot ID (if applicable)

Search Conversations

Query Parameters

query
string
required
Search keyword or phrase

Response

id
string
Conversation identifier
title
string
Conversation title
last_updated_time
float
Unix timestamp of last update
bot_id
string
Bot ID (if applicable)
highlights
array
Search highlights showing matching text fragments

Delete Conversation

Path Parameters

conversation_id
string
required
The conversation ID to delete

Delete All Conversations


Bulk Delete Conversations

Request Body

conversation_ids
array
required
Array of conversation IDs to delete (1-100 items)

Response

deleted_count
integer
Number of successfully deleted conversations
failed_ids
array
Array of conversation IDs that failed to delete

Update Conversation Title

Path Parameters

conversation_id
string
required
The conversation ID to update

Request Body

new_title
string
required
The new title for the conversation

Propose Conversation Title

Path Parameters

conversation_id
string
required
The conversation ID

Response

title
string
Suggested conversation title

Submit Feedback

Path Parameters

conversation_id
string
required
The conversation ID
message_id
string
required
The message ID to provide feedback on

Request Body

thumbs_up
boolean
required
True for positive feedback, false for negative
category
string
Reason category (required if thumbs_up is false)
comment
string
Optional detailed comment

Response

thumbs_up
boolean
The feedback value
category
string
The category
comment
string
The comment

Path Parameters

conversation_id
string
required
The conversation ID

Response

Returns an array of related documents:
content
object
Tool result content (text, JSON, image, or document)
source_id
string
Unique identifier for the source
source_name
string
Name of the source document
Link to the source (S3 URL or HTTP URL)
page_number
integer
Page number in the source document
citation_number
integer
Sequential citation number for inline references
excerpt
string
Short text excerpt for citation preview
relevance_score
float
Relevance score from 0.0 to 1.0
freshness_date
string
Date the source was last updated (YYYY-MM-DD)
quality_score
float
Combined quality score from 0.0 to 1.0

Path Parameters

conversation_id
string
required
The conversation ID
source_id
string
required
The source document ID

Get Document Preview

Path Parameters

conversation_id
string
required
The conversation ID
source_id
string
required
The source document ID

Response

source_id
string
The source ID of the document
source_name
string
The name of the source
content_type
string
Content type: "text", "pdf", "image", or "binary"
preview_url
string
Presigned URL for viewing binary files (valid for 15 minutes)
text_content
string
Text content for text-based files (first 10,000 characters)
page_count
integer
Number of pages for paginated documents
file_size
integer
File size in bytes

Build docs developers (and LLMs) love