Skip to content
/
Assistants Overview
Last updated

Assistants

Assistants are AI-powered agents in Freddy that can be configured with specific behaviors, knowledge, and tools to help users accomplish tasks.

Quick Overview

An assistant combines:

  • AI Model: GPT-4, Claude, or other supported models
  • Instructions: System prompts defining behavior
  • Knowledge: Rules, documents, and context
  • Tools: System tools, MCP integrations, and automations
  • Configuration: Reasoning, output format, and context settings

Key Features

  • Purpose-Built: Create specialized agents for different tasks
  • Intelligent Reasoning: Configure chain-of-thought and multi-step planning
  • Tool Integration: System tools, MCP tools, and Streamline automations
  • Knowledge Management: Attach rules, vector stores, and documents
  • Access Control: Organization and user-level permissions

Database Architecture

Assistants use a normalized database structure with 7 related tables:

  1. assistants - Main table with core assistant data
  2. assistant_tool_configurations - Tools (system, MCP, Streamline)
  3. assistant_rules - Knowledge and guidelines
  4. assistant_reasoning - Thinking configuration
  5. assistant_json_schemas - Structured output formats
  6. assistant_vector_stores - Semantic search connections
  7. assistant_rule_attachments - File attachments

This design provides flexibility, data integrity, and efficient querying.

Quick Start

Create an Assistant

curl https://api.aitronos.com/v1/assistants \
 -H "X-API-Key: $FREDDY_API_KEY" \
 -H "Content-Type: application/json" \
 -d '{
 "name": "Support Agent",
 "instructions": "You are a helpful customer support agent.",
 "model": "gpt-4o"
 }'

Configure Tools

curl https://api.aitronos.com/v1/assistants/{assistant_id}/tools \
 -H "X-API-Key: $FREDDY_API_KEY" \
 -d '{
 "system_tools": {
 "file_search": {"enabled": true},
 "web_search": {"enabled": true}
 },
 "mcp_tools": ["mcp_abc123"]
 }'

Use the Assistant

curl https://api.aitronos.com/v1/model/response \
 -H "X-API-Key: $FREDDY_API_KEY" \
 -H "Content-Type: application/json" \
 -d '{
 "assistant_id": "asst_abc123",
 "thread_id": "thrd_xyz789",
 "inputs": [{"role": "user", "content": "Hello!"}]
 }'

Detailed Documentation

For comprehensive guides, see the dedicated assistants documentation:

Common Use Cases

Customer Support

{
 "name": "Support Agent",
 "instructions": "You are a helpful customer support agent...",
 "tools": {
 "system_tools": ["file_search", "web_search"],
 "mcp_tools": ["zendesk_integration"]
 }
}

Code Review

{
 "name": "Code Reviewer",
 "instructions": "You are an expert code reviewer...",
 "tools": {
 "system_tools": ["code_interpreter"],
 "mcp_tools": ["github_integration"]
 },
 "reasoning": {"enabled": true}
}

Data Analysis

{
 "name": "Data Analyst",
 "instructions": "You are a data analysis expert...",
 "tools": {
 "system_tools": ["code_interpreter"],
 "mcp_tools": ["database_connector"]
 }
}
Previous page
Next page