agentwatch is a platform-agnostic AI observability framework that provides real-time insights and detailed tracking of AI agent interactions to optimize AI-driven applications.
A powerful AI observability framework that provides comprehensive insights into agent interactions across platforms, enabling developers to monitor, analyze, and optimize AI-driven applications with minimal integration effort.
Developers and AI practitioners use agentwatch to monitor, analyze, and optimize interactions between AI agents and tools across multiple frameworks, enabling improved debugging, performance tuning, and security oversight. It is ideal for those building or maintaining AI-driven applications who need comprehensive visibility into agent behavior with minimal integration effort.
agentwatch requires Python 3.11+ and npm for UI visualization; it currently supports HTTP/SSE for remote MCP server tracking and is designed to be framework-agnostic but primarily targets Python AI applications. Users should place their main execution code inside a __main__ block to ensure proper instrumentation.
Ensure Python 3.11+ and npm are installed
Install agentwatch via pip from GitHub: pip install git+https://github.com/cyberark/agentwatch.git
Clone the repository to access examples and additional resources
Install python dependencies for examples using Poetry: poetry install --with dev
Run the UI locally using the provided script: ./run_ui.shimport agentwatch
Import agentwatch in your main Python module to enable observability.
if __name__ == "__main__":
Ensure your main code execution is within this block for proper tracking.
agentwatch ui
Launch the agentwatch UI which opens a new browser tab to visualize AI agent interactions.
poetry run python examples/langgraph_example.py
Run example scripts to see agentwatch in action with sample AI frameworks.