Skip to main content

​
What is an Agent?

After deploying an agent to AgentOS, it appears in the Agents List (/agents/) with an Online status. Click the Manage button to access the agent’s detail page where you can monitor, test, and manage your agent. Each deployed agent gets:
  • Dedicated API endpoint (IP:port)
  • Automatic OpenAPI documentation
  • Execution history tracking
  • Performance metrics
  • Scheduling capabilities

​
Management Page

The agent management page provides complete control and visibility into your agent’s operations.

​
API Endpoint

Each agent is assigned a unique IP and port. For local environments:
  • First agent: http://0.0.0.0:7340
  • Subsequent agents: Sequential ports (7341, 7342, etc.)

​
Making API Calls

Call your agent using the /call endpoint: 
curl -X POST "http://0.0.0.0:7340/call" \
  -F "Mail=example_Mail"
Get the result using the job ID: 
curl "http://0.0.0.0:7340/result/JOB_ID"

​
What You’ll See

The management page displays:
  • Health Status - API and Web UI health checks
  • API Endpoint - Direct link to /docs for OpenAPI documentation
  • Example Code - cURL, Python, JavaScript, and other language examples
  • Streamlit UI - Access to the Streamlit interface (if available)
  • Test Endpoints - Interactive API testing interface
Agent Management Page

​
OpenAPI Documentation

Every agent automatically generates OpenAPI documentation. Access it at: 
http://IP:PORT/docs
The documentation provides:
  • Interactive API testing interface
  • Request/response schemas
  • Authentication details
  • Example requests in multiple languages
OpenAPI Documentation

​
Executions

View the complete execution history of your agent, including:
  • Timestamp of each execution
  • Input parameters
  • Output results
  • Execution duration
  • Success/failure status
Filter and search through executions to analyze your agent’s performance over time.
Execution History

​
Benchmark

Compare your agent’s performance across different LLM models to find the optimal configuration.

​
How It Works

  1. Select multiple LLM connections (configured in Settings)
  2. Provide the same input parameters
  3. AgentOS calls your agent with each model
  4. Compare results side-by-side

​
What You Can Compare

  • Cost - API costs for each model
  • Response Time - Inference speed comparison
  • Token Usage - Token consumption analysis
  • Output Quality - Manual comparison of responses
This helps you choose the most cost-effective and performant model for your use case.
Benchmark Comparison

​
Results View

After running the benchmark, view detailed comparison results for each model:
Benchmark Results Comparison

​
Metrics

Monitor your agent’s performance with detailed metrics and visualizations. AgentOS automatically tracks:
  • Total Calls - Number of executions
  • Total Cost - Cumulative API costs
  • Average Response Time - Performance trends
  • Token Usage - Input and output tokens
View both overall statistics and time-series graphs to identify trends and optimize performance. For detailed information about metrics features and the platform-wide Usage page, see the Metrics guide.
Agent Metrics Dashboard

​
Scheduling

Automate agent execution with flexible scheduling options. Configure your agent to run automatically at specific times or intervals using:
  • Frequency-Based - Hourly, daily, or weekly execution
  • Crontab Expression - Custom schedules with cron syntax
For detailed scheduling configuration and examples, see the Scheduling guide.
Agent Scheduling

​
Deployments

Track your agent’s deployment history and monitor deployment status.

​
Deployment History

View all deployments with:
  • Deployment Status - Success or failure
  • Timestamp - When the deployment occurred
  • Commit Hash - Git commit used for deployment (helps identify which code version is running)
  • Failure Reasons - Detailed error messages for failed deployments
This helps you track changes and understand which version of your code is currently deployed.
Deployment History