Chat with Agent

POST /v1/aibot/bots/{AgentId}/send-message

Chat with Agent

cURL

curl --request POST \
  --url http://{your-domain}/v1/aibot/bots/{AgentId}/send-message \
  --header 'Authorization: Bearer <token>' \
  --header 'Content-Type: application/json' \
  --data '
{
  "messages": [
    {
      "content": "<string>"
    }
  ],
  "tools": [
    {
      "name": "<string>",
      "description": "<string>",
      "parameters": {}
    }
  ],
  "conversationId": "conversation-xxx"
}
'

200

{
  "content": "Hello! How can I assist you today?"
}

Authorizations

Authorization · string · header · required

Token corresponding to the environment ID, obtained through login authentication

Headers

Accept · string · default: text/event-stream

Accept header for streaming response

Path Parameters

AgentId · string · required

Agent ID

Body

application/json

messages · object[]

Array of conversation messages

Show child attributes

tools · object[]

Array of available tools for the agent

Show child attributes

resume · object

Resume info for resuming an interrupted conversation

Show child attributes

conversationId · string

Conversation ID for multi-conversation support (optional)

Example:

"conversation-xxx"

Response

200

text/event-stream

Agent conversation successful, returns SSE format. Each event follows the SSE specification with JSON data:

data: {"type": "text", "content": "Hello"}

data: {"type": "tool-call-start", "toolCallId": "call_123", "toolCallName": "search"}

data: {"type": "tool-call-args", "toolCallId": "call_123", "delta": "{\"query\": \"weather\"}"}

data: {"type": "tool-call-end", "toolCallId": "call_123"}

data: {"type": "tool-result", "result": "Sunny, 25°C", "toolCallId": "call_123"}

data: [DONE]

The response is a stream of Server-Sent Events where each event contains JSON data representing different types of agent interactions.

Text content event from the agent. Contains streaming text responses as the agent generates them.

type · enum<string> · required

Event type identifier

Available options: text

content · string · required

Text content from agent

Example:

"Hello! How can I assist you today?"