MCP Server Agentops

An MCP server project that provides an interface for LLM agents to interact with the AgentOps API for accessing tracking data, including authentication processes and tool call functions.

Developer tools Monitoring #API integration #Data tracking #Agent tools .Python

rating : 2 points

downloads : 11

update time : 2025-04-28

What is the AgentOps API MCP Server?

This is a server specifically designed for LLM (Large Language Model) agents, allowing them to securely access and query tracking data on the AgentOps platform. It simplifies the interaction process between agents and the AgentOps API.

How to use the AgentOps API MCP Server?

You can use it through the Claude Coder tool or by directly running the server. The main functions include viewing the tracking list and getting tracking details.

Use cases

When you need to analyze the running performance of LLM agents, debug issues, or optimize workflows, this server can help you easily obtain relevant tracking data.

Main features

List tracking recordsGet a list of recent tracking records and specify the number of records to return.

Tracking detailsGet detailed information about a specific tracking, including all relevant data points.

Secure authenticationAdopt a two-step authentication process to ensure the security of API keys.

Advantages and limitations

Advantages

Simplify the process for LLM agents to access AgentOps data

Adopt a secure JWT token authentication mechanism

Support custom API URLs for easy development and testing

Seamlessly integrate with Claude Coder

Limitations

Currently only support two basic query functions

Require providing an API key for each request

Do not support advanced filtering and search functions for now

How to use

Configure Claude Coder

Add the MCP server to the Claude Coder configuration file.

Run Claude Coder

Start Claude Coder and specify to use the AgentOps MCP server.

Query tracking data

Include your AgentOps API key in the prompt to query data.

Usage examples

Performance analysisAnalyze the execution time and resource usage of the last 5 traces.

Error debuggingCheck error information in a specific trace.

Frequently Asked Questions

How to get an AgentOps API key?

Can it be used in a local development environment?

What is the format of the returned data?

Related resources

AgentOps official documentation

Complete documentation for the AgentOps platform

Claude Coder GitHub repository

Source code and documentation for Claude Coder

MCP protocol specification

Technical specification for the Model Coordination Protocol

🚀 AgentOps API MCP Server

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

🚀 Quick Start

The AgentOps API MCP Server is designed to enable LLM agents to access trace data via the AgentOps API. It offers functionalities such as listing recent traces and retrieving detailed information about specific traces.

✨ Features

Overview

This MCP server provides agents with the following capabilities:

List recent traces in their AgentOps projects.
Obtain detailed information about specific traces.

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

Authentication Process

The server uses a secure two-step authentication process:

The agent provides its AgentOps API key with each request.
The server exchanges this API key for a JWT token obtained from the AgentOps API.
The server uses this JWT token to make authenticated requests to the AgentOps API.
All responses are returned to the agent in their original form.

This ensures that the API key is never stored between requests and that each request is securely authenticated using 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 IDs, the number of spans, start/end times.
The total number of traces in the database.
Sorted by creation timestamp (in descending order).

`trace_detail`

Get detailed information about a specific trace. Parameters:

AGENTOPS_API_KEY (required): The agent's AgentOps API key.
trace_id (required): The trace ID for which to retrieve detailed information (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

Use with Claude Coder (Recommended)

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

Configure Claude Coder Add the following content to your Claude Coder configuration file (usually ~/.config/claude-cli/config.yaml):

mcp_servers:
  - name: agentops-api
    path: /ABSOLUTE/PATH/TO/mcp/agentops-api/bin/run-server
    description: "AgentOps API integration to access trace data"

Replace /ABSOLUTE/PATH/TO with the actual path to this repository.

Use with Claude Coder

# Start Claude and enable the AgentOps API MCP server
claude --mcp agentops-api

# Or for a single command
claude --mcp agentops-api "List my recent traces. Please find my agent API key in my user .env file."

Include your AgentOps API key in your prompt to Claude:

Can you show me my recent traces? Please find my agent API key in my user .env file.

Run as a Stand - Alone Server

If you need to run the server directly:

# Use the convenience script
./bin/run-server

# Or use Python directly
uv run -m mcp_server_agentops_api

The server uses the production AgentOps API https://api.agentops.ai by default. Agents can override this URL with each request.

Available Tool Calls

When using the MCP server with Claude, the following tools can be used:

# List recent traces
list_traces(AGENTOPS_API_KEY="your-api-key", limit=10)

# Get detailed information about a specific trace
trace_detail(AGENTOPS_API_KEY="your-api-key", trace_id="148dac266d95c9dc0b5616320b8488c9")

# Use a custom API URL (e.g., for local development)
list_traces(AGENTOPS_API_KEY="your-api-key", AGENTOPS_API_URL="http://localhost:8080")