By comet-mlCreated 16 days ago
starstarstarstarstar

Query and analyze your Opik logs, traces, prompts and other LLM telemetry in natural language.

Visit Project
Share this MCP:
X (Formerly Twitter)RedditblueskyThreads by Instagram

Category

Official MCP Server

Tags

OpikLlm TelemetryMcpData Analysis

what is Comet Opik?

Comet Opik is an open-source engine that allows users to query and analyze their Opik logs, traces, prompts, and other LLM telemetry in natural language through an implementation of the Model Context Protocol (MCP). It provides a unified interface for interacting with Opik's capabilities and supports multiple transport mechanisms for integration with various environments, including IDEs.

How to use Comet Opik?

To use Comet Opik:

  1. Clone the repository (git clone https://github.com/comet-ml/opik-mcp.git).
  2. Install dependencies and build the project (npm install && npm run build).
  3. Configure the server by creating a .env file.
  4. Start the server with either stdio (npm run start:stdio) or SSE transport (npm run start:sse, experimental).
  5. Integrate with IDEs like Cursor by setting up a .cursor/mcp.json file.

Key features of Comet Opik?

  • Prompts Management: Create, list, update, and delete prompts.
  • Projects/Workspaces Management: Organize and manage projects/workspaces.
  • Traces: Track and analyze LLM execution traces.
  • Metrics: Query and analyze metrics data.
  • Transport Options: Supports stdio (local) and SSE (experimental, network-based).
  • IDE Integration: Seamlessly works with Cursor and compatible IDEs.

Use cases of Comet Opik?

  1. Debugging LLM Applications: Analyze traces and metrics to identify issues.
  2. Workflow Automation: Integrate prompts and projects into development environments.
  3. Multi-Client Environments: Use SSE transport for remote/parallel interactions.
  4. Research & Prototyping: Experiment with LLM telemetry via natural language queries.

FAQ from Comet Opik?

  • Q: What is the MCP protocol?

    A: The Model Context Protocol (MCP) standardizes communication between LLM tools and clients (e.g., IDEs).

  • Q: Is the SSE transport stable for production?

    A: No, SSE is experimental. Use stdio for production deployments.

  • Q: How do I integrate with Cursor IDE?

    A: Configure .cursor/mcp.json with the server path and your Opik API key.

  • Q: Where can I find documentation?

    A: See the official docs for guides on API reference, transports, and IDE setup.