MCP.so
Sign In

mcp-server-agentops

@AgentOps-AI

About mcp-server-agentops

Official MCP server for the AgentOps SDK

Basic information

Category

Other

License

Apache-2.0

Runtime

python

Transports

stdio

Publisher

AgentOps-AI

Config

Add this server to your MCP-compatible client using the configuration below.

{
  "mcpServers": {
    "mcp-server-agentops": {
      "command": "uv",
      "args": [
        "run",
        "-m",
        "mcp_server_agentops_api"
      ]
    }
  }
}

Tools

No tools detected

We auto-extract tools from the README. The maintainer can list them under a ## Tools heading to populate this section.

Overview

What is mcp-server-agentops?

mcp-server-agentops is an MCP (Model Context Protocol) server that enables LLM agents to read trace data from the AgentOps API. It provides tools to list recent traces and retrieve detailed information about specific traces, handling authentication and API communication on behalf of the agent.

How to use mcp-server-agentops?

The server can be run interactively through Claude Coder (recommended) or as a standalone process. When using Claude Coder, add the server to your Claude configuration file (~/.config/claude-cli/config.yaml) with the correct path to the run-server script, then start Claude with --mcp agentops-api. Agents include their AgentOps API key in each prompt or tool call. Standalone execution is possible via ./bin/run-server or uv run -m mcp_server_agentops_api.

Key features of mcp-server-agentops

  • List recent traces from an AgentOps project
  • Get detailed information about a specific trace
  • Secure two-step authentication (API key → JWT token)
  • Stateless operation; no API keys stored between requests
  • Customizable API URL for non‑production environments
  • Designed for use with Claude Coder

Use cases of mcp-server-agentops

  • An agent reviews recent traces to monitor its own execution history
  • A developer fetches detailed trace data for debugging a specific agent run
  • An automated workflow checks trace counts to verify system health
  • A non‑production environment uses a custom API URL to test against staging

FAQ from mcp-server-agentops

What is required to use mcp-server-agentops?

An AgentOps API key is mandatory for every request. The key is passed as a parameter to the list_traces or trace_detail tool.

How does authentication work?

The server exchanges the provided API key for a JWT token from the AgentOps API, then uses that token for subsequent authenticated requests. The API key is never stored between requests.

What tools does mcp-server-agentops provide?

Two tools: list_traces (returns recent traces with IDs, spans, and timestamps) and trace_detail (returns full details for a given trace ID).

Can I use a custom API URL?

Yes. Both tools accept an optional AGENTOPS_API_URL parameter. The default is the production endpoint at https://api.agentops.ai.

Does the server store any data between requests?

No. Each request is stateless; raw API responses are returned directly to the agent, and no user data is persisted.

Comments

More Other MCP servers