Empower Your AI Applications with Unparalleled Observability and Optimization
AgentNeo is an advanced, open-source Agentic AI Application Observability, Monitoring, and Evaluation Framework. Designed to elevate your AI development experience, AgentNeo provides deep insights into your AI agents, Large Language Model (LLM) calls, and tool interactions. By leveraging AgentNeo, you can build more efficient, cost-effective, and high-quality AI-driven solutions.
Whether you're a seasoned AI developer or just starting out, AgentNeo offers robust logging, visualization, and evaluation capabilities to help you debug and optimize your applications with ease.
Install AgentNeo effortlessly using pip:
pip install agentneo
Get up and running with AgentNeo in just a few steps!
from agentneo import AgentNeo, Tracer, Evaluation, launch_dashboard, Execution
neo_session = AgentNeo(session_name="my_session")
neo_session.create_project(project_name="my_project")
tracer = Tracer(session=neo_session)
tracer.start()
Wrap your functions with AgentNeo's decorators to start tracing:
@tracer.trace_llm("my_llm_call")
async def my_llm_function():
# Your LLM call here
pass
@tracer.trace_tool("my_tool")
def my_tool_function():
# Your tool logic here
pass
@tracer.trace_agent("my_agent")
def my_agent_function():
# Your agent logic here
pass
exe = Execution(session=neo_session, trace_id=1)
# run a single metric
exe.execute(metric_list=['metric_name'])
# get your evaluated metrics results
metric_results = exe.get_results()
print(metric_results)
tracer.stop()
launch_dashboard(port=3000)
Access the interactive dashboard by visiting http://localhost:3000
in your web browser.
Manage multiple projects with ease.
List All Projects
projects = neo_session.list_projects()
Connect to an Existing Project
neo_session.connect_project(project_name="existing_project")
Run multiple metrics together
exe.execute(metric_list=['metric_name1', 'metric_name2', ..])
Use your own config and metadata related to the metric
exe.execute(metric_list=['metric_name'], config={}, metadata={})
AgentNeo generates an execution graph that visualizes the flow of your AI application, including LLM calls, tool usage, and agent interactions. Explore this graph in the interactive dashboard to gain deeper insights.
The AgentNeo dashboard offers a comprehensive view of your AI application's performance:
neo_session.launch_dashboard(port=3000)
We are committed to continuously improving AgentNeo. Here's a glimpse of what's on the horizon:
Feature | Status |
---|---|
Local Data Storage Improvements | ā Completed |
Support for Additional LLMs | ā Completed |
Integration with AutoGen | ā Completed |
Integration with CrewAI | ā Completed |
Integration with Langraph | ā Completed |
Tracing User Interactions | ā Completed |
Tracing Network Calls | ā Completed |
Comprehensive Logging Enhancements | ā Completed |
Custom Agent Orchestration Support | ā Completed |
Advanced Error Detection Tools | š In Progress |
Multi-Agent Framework Visualization | ā Completed |
Performance Bottleneck Identification | ā Completed |
Evaluation Metrics for Agentic Application | ā Completed |
Code Execution Sandbox | š Coming Soon |
Prompt Caching for Latency Reduction | š Planned |
Real-Time Guardrails Implementation | š Planned |
Open-Source Agentic Apps Integration | š Planned |
Security Checks and Jailbreak Detection | š Planned |
Regression Testing Capabilities | š Planned |
Agent Battleground for A/B Testing | š Planned |
IDE Plugins Development | š Planned |
For more details, explore the full AgentNeo Documentation
We warmly welcome contributions from the community! Whether it's reporting bugs, suggesting new features, or improving documentation, your input is invaluable.
Join us in making AgentNeo even better!