Agentops MCP Server

🚀 AgentOps API MCP Server

A Model Coordination Protocol (MCP) server that allows LLM agents to access trace data via the AgentOps API.

🚀 Quick Start

This MCP server enables agents to perform the following functions:

• List recent traces in a project.
• Retrieve detailed information about a specific trace.

The server handles authentication and communication with the AgentOps API, allowing agents to easily access their trace data.

✨ Features

Authentication Process

The server employs a secure two - step authentication process:

1. The agent provides its AgentOps API key with each request.
2. The server exchanges this API key with the AgentOps API for a JWT token.
3. The server uses this JWT token to authenticate and access the AgentOps API.
4. All responses are returned to the agent in their original form.

This ensures that the API key is never stored between requests, and each request is securely authenticated via the JWT token.

Available Tools

list_traces

List recent traces for an agent's project.

Parameters:

• AGENTOPS_API_KEY (required): The agent's AgentOps API key.
• limit (optional): The maximum number of traces to return (default: 5).
• AGENTOPS_API_URL (optional): A custom API URL for non - production environments.

Response:

• A list of traces including trace ID, number of spans, start/end times.
• The total number of traces in the database.
• Sorted by creation timestamp (from most recent to oldest).

trace_detail

Retrieve detailed information about a specific trace.

Parameters:

• AGENTOPS_API_KEY (required): The agent's AgentOps API key.
• trace_id (required): The ID of the trace for which to retrieve details (obtained from the trace_id field in the list_traces response).
• AGENTOPS_API_URL (optional): A custom API URL for non - production environments.

Response:

• Detailed information about the trace.
• All spans associated with the trace.
• Metadata, time information, etc.

💻 Usage Examples

Using with Claude Coder (Recommended)

This MCP server is designed to work with Claude Coder. Follow these steps to configure it:

Basic Usage

# Add the following to your Claude Coder configuration file (usually ~/.config/claude/config.yml)
agents:
  - type: mcp_server
    name: AgentOps MCP Server
    config:
      server_url: "http://localhost:8000"

Advanced Usage

# Run the following command in your terminal to start the MCP server
python3 -m mcp_server --host localhost --port 8000

🔧 Technical Details

Architecture

The system consists of three main components:

• Client: Responsible for interacting with agents and receiving requests.
• Tools: Contains functional modules such as list_traces and trace_detail.
• Server: Handles authentication, communication, and request processing.

Future Enhancements

• Add more AgentOps API tools (metrics, spans, etc.).
• Support filtering and searching for traces.
• Implement batch operations to improve performance.
• Cache JWT tokens for short - term optimization.