Query Stream

Used to send queries to an AI agent and receive streaming responses. This endpoint allows you to interact with an AI agent's chat history for a specific project and receive real-time responses in a streaming format.

API Endpoint

Property	Value
Request Method	POST
Request URL	https://api.seliseblocks.com/ai-agent/query/stream

Request

Request Example

curl -X POST 'https://api.seliseblocks.com/ai-agent/query/stream' \
  -H 'accept: application/json' \
  -H 'Content-Type: application/json' \
  -d '{
    "query": "What are the features of this product?",
    "session_id": "session_123456",
    "base_prompt": "You are a helpful customer support assistant",
    "model_id": "gpt-4",
    "model_name": "GPT-4",
    "model_provider": "openai",
    "tool_ids": ["tool_1", "tool_2"],
    "last_n_turn": 5,
    "enable_summary": false,
    "enable_next_suggestion": false,
    "response_type": "text",
    "response_format": "markdown",
    "call_from": "project_slug_123",
    "files": ["file_id_1"],
    "images": ["image_url_1"]
  }'

Request Headers

Field	Type	Required	Description
accept	string	Yes	Accepted response format. Use application/json
Content-Type	application/json	Yes	Data type, must be application/json.

Request Body

Request Body Schema

{
  "query": "string",
  "session_id": "string",
  "base_prompt": "string",
  "model_id": "string",
  "model_name": "string",
  "model_provider": "string",
  "tool_ids": ["string"],
  "last_n_turn": 5,
  "enable_summary": false,
  "enable_next_suggestion": false,
  "response_type": "text",
  "response_format": "string",
  "call_from": "string",
  "files": ["string"],
  "images": ["string"]
}

Request Body Parameters

Field	Type	Required	Description
query	string	Yes	The user's query or message to send to the AI agent.
session_id	string	No	Unique identifier for the chat session. Used to maintain conversation context.
base_prompt	string	No	System prompt or base instruction for the AI agent to follow.
model_id	string	No	Identifier of the language model to use.
model_name	string	No	Human-readable name of the model (e.g., GPT-4, Claude).
model_provider	string	No	Provider of the model (e.g., openai, anthropic).
tool_ids	array	No	Array of tool identifiers that the agent can utilize to answer the query.
last_n_turn	integer	No	Number of previous conversation turns to include as context. Default: 5.
enable_summary	boolean	No	Whether to enable automatic summarization of the conversation. Default: false.
enable_next_suggestion	boolean	No	Whether to enable suggestions for follow-up queries. Default: false.
response_type	string	No	Format of the response (e.g., text, json). Default: text.
response_format	string	No	Specific formatting instructions for the response (e.g., markdown, html).
call_from	string	No	Project slug identifier to specify which project's AI agent to query. Required to retrieve project-specific chat history.
files	array	No	Array of file identifiers to include as context for the query.
images	array	No	Array of image URLs to include as context for the query.

note

• The call_from parameter is used to identify which project's AI agent should handle the query and retrieve the corresponding chat history.
• The session_id helps maintain conversation continuity across multiple queries.
• The last_n_turn parameter controls how much previous conversation context is provided to the model.

Response

Success Response (200 OK)

Returns a streaming text response containing the AI agent's reply.

"The product includes the following features: 1. Real-time chat capabilities, 2. Integration with multiple AI models, 3. Session management, 4. Context-aware responses, and 5. Support for file and image attachments. Would you like more details on any specific feature?"

Response Description

Field	Type	Description
(stream)	string	The response is streamed as text. The AI agent's reply to the query, which may include contextual information from the chat history and available tools. The stream continues until the complete response is transmitted.

info

The response is returned as a stream. This means the data is sent incrementally as it's generated, allowing for real-time display of the AI agent's response.

Error Response

422 Validation Error

Returned when the request body contains invalid or missing required fields.

{
  "detail": [
    {
      "loc": ["query"],
      "msg": "Field required",
      "type": "missing",
      "input": null,
      "ctx": {}
    }
  ]
}

Error Fields

Field	Type	Description
detail	array	Array of validation error objects.
loc	array	Location of the error (field name).
msg	string	Human-readable error message.
type	string	Type of validation error.
input	any	The input value that caused the error.

Use Cases

• Customer Support: Send customer inquiries and retrieve AI-generated responses from a support agent.
• Knowledge Base Queries: Query an AI agent to retrieve information from project-specific knowledge bases.
• Multi-turn Conversations: Maintain context across multiple queries using the session_id.
• Tool Integration: Leverage AI agent tools to provide enhanced responses with relevant data.
• File & Image Analysis: Include files and images as context for more accurate and detailed responses.

Example Use Case

Retrieving chat history for a project's AI agent:

curl -X POST 'https://api.seliseblocks.com/ai-agent/query/stream' \
  -H 'accept: application/json' \
  -H 'Content-Type: application/json' \
  -d '{
    "query": "Show me the order history for customer ABC123",
    "session_id": "session_user_001",
    "call_from": "ecommerce_project",
    "model_name": "GPT-4",
    "model_provider": "openai",
    "tool_ids": ["order_database_tool", "customer_lookup_tool"],
    "last_n_turn": 10,
    "enable_summary": true
  }'

This request sends a query to the AI agent associated with the ecommerce_project, utilizing tools to fetch order data and maintain context of the last 10 conversation turns.