Skip to main content

OverseeX integration for Microsoft AutoGen - automatic trace capture for multi-agent conversations

Project description

OverseeX AutoGen Integration

Auto-instrumentation for Microsoft AutoGen multi-agent conversations.

Installation

pip install overseex-autogen

Quick Start

from autogen import AssistantAgent, UserProxyAgent
from overseex_autogen import monitor_autogen

# Create your AutoGen agents
assistant = AssistantAgent("assistant", llm_config=llm_config)
user_proxy = UserProxyAgent("user_proxy")

# Monitor the conversation
with monitor_autogen(api_key="ag_live_...", agents=[assistant, user_proxy]) as monitor:
    user_proxy.initiate_chat(assistant, message="What is the capital of France?")

# All interactions are automatically traced to OverseeX!

Features

  • Zero-config Auto-instrumentation: Just wrap your code and traces are captured
  • Multi-turn Conversation Tracking: Every message between agents is recorded
  • Function Call Monitoring: Track all function/tool executions
  • Agent Role Analysis: Monitor agent behaviors and coordination
  • Coordination Intelligence: Detect handoff patterns and issues

Usage

Context Manager

from overseex_autogen import AutoGenMonitor

with AutoGenMonitor(api_key="your-api-key") as monitor:
    # Start tracking
    monitor.start_chat([assistant, user_proxy])

    # Your AutoGen code
    result = user_proxy.initiate_chat(assistant, message="Hello!")

    # End tracking
    monitor.end_chat()

Manual Callback

from overseex_autogen import OverseeXAutoGenCallback

callback = OverseeXAutoGenCallback(
    api_key="your-api-key",
    capture_functions=True,
    capture_messages=True,
    verbose=True
)

# Register agents
callback.register_agent(assistant)
callback.register_agent(user_proxy)

# Start conversation tracking
callback.start_conversation([assistant, user_proxy])

# ... run your AutoGen code ...

# End and send traces
callback.end_conversation()

What Gets Captured

  • Messages: All inter-agent communication
  • Function Calls: Tool usage with inputs/outputs
  • Agent Roles: AssistantAgent, UserProxyAgent, GroupChat, etc.
  • Conversation Flow: Turn-by-turn execution
  • Token Usage: LLM token consumption
  • Coordination Events: Handoffs between agents

Configuration

OverseeXAutoGenCallback(
    api_key="your-api-key",           # Required
    base_url="https://api.overseex.com",  # Optional
    capture_functions=True,           # Capture function calls
    capture_messages=True,            # Capture message content
    capture_system_prompts=False,     # Capture system prompts (may contain sensitive data)
    auto_register_agents=True,        # Auto-register agents in OverseeX
    max_message_length=2000,          # Truncate long messages
    verbose=False,                    # Enable debug logging
)

License

MIT

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

overseex_autogen-0.1.0.tar.gz (9.1 kB view details)

Uploaded Source

Built Distribution

If you're not sure about the file name format, learn more about wheel file names.

overseex_autogen-0.1.0-py3-none-any.whl (8.5 kB view details)

Uploaded Python 3

File details

Details for the file overseex_autogen-0.1.0.tar.gz.

File metadata

  • Download URL: overseex_autogen-0.1.0.tar.gz
  • Upload date:
  • Size: 9.1 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/6.2.0 CPython/3.14.2

File hashes

Hashes for overseex_autogen-0.1.0.tar.gz
Algorithm Hash digest
SHA256 c1fd65d86c68c2a6c03ebc734ec0268b583d5d7383200ebd85bfebe99eb81e23
MD5 fad4929e87f24062c3a3c77b670251dc
BLAKE2b-256 24f31483d74cb43b7b0b246cbec5047a220c767056dc9a9477fcb6603606b3ef

See more details on using hashes here.

File details

Details for the file overseex_autogen-0.1.0-py3-none-any.whl.

File metadata

File hashes

Hashes for overseex_autogen-0.1.0-py3-none-any.whl
Algorithm Hash digest
SHA256 51c7e158b81a276d9ef8d63263f673159f439eecc63d5ee9f5e961129cbca864
MD5 15a740a73081325b448942ac1e425d42
BLAKE2b-256 85f326f6d1fd7cd6d87f3e22fc5b784371731eb89013208da46ea2755728ee27

See more details on using hashes here.

Supported by

AWS Cloud computing and Security Sponsor Datadog Monitoring Depot Continuous Integration Fastly CDN Google Download Analytics Pingdom Monitoring Sentry Error logging StatusPage Status page