Skip to main content

BizOSaaS Platform — Agent Configuration Guide

This guide explains how to configure and fine-tune the AI agents within the BizOSaaS ecosystem.

1. Core Agent Structure​

All agents inherit from BaseAgent and follow the Hexagonal Architecture pattern.

  • Port: Defined in app/ports/ai_port.py
  • Adapter: Located in ai-agents/agents/

2. Environment Configuration​

Agents use the following environment variables (managed in Vault):

  • OPENROUTER_API_KEY: Primary LLM provider.
  • SERP_API_KEY: Used by research agents.
  • AGENT_VERBOSE: Set to true for detailed logging.

3. Fine-Tuning Performance​

  • Temperature: Adjust in the agent initialization to control creativity (0.0 for deterministic, 0.7 for creative).
  • Context Window: Configurable per agent to handle large documents.
  • Tools: Agents can be granted access to specific connectors via the ConnectorRegistry.

4. Registering New Agents​

  1. Create your agent class in ai-agents/agents/.
  2. Inherit from BaseAgent.
  3. Override _execute_agent_logic.
  4. Register the agent in the worker.py for Temporal workflow access.