Comet-ML/Opik-MCP
Use natural language to explore LLM observability, traces, and monitoring data captured by Opik.
Ads
Overview
Use natural language to explore LLM observability, traces, and monitoring data captured by Opik.
Installation
claude mcp add --transport stdio opik-mcp --env OPIK_API_KEY=<your-key> --env OPIK_WORKSPACE=<your-workspace> -- uvx opik-mcp
Configuration
{
"mcpServers": {
"opik-mcp": {
"type": "stdio",
"command": "uvx",
"args": ["opik-mcp"],
"env": {
"OPIK_API_KEY": "<your-key>",
"OPIK_WORKSPACE": "<your-workspace>"
}
}Capabilities
- Read any Opik entity
Fetch projects, traces, spans, test suites, experiments, or prompts by id, name, or `opik://` URI via `read`.
- Browse collections
List entities like experiments or project-scoped traces with name filters and pagination using `list`.
- Ask Ollie investigative questions
Send cross-entity questions about latency, regressions, or comparisons to `ask_ollie` for synthesis.
- Log scores and comments
Attach numeric feedback scores or free-text comments to traces, spans, or threads via `write` operations like `score.create`.
- Save prompt versions
Create or update a prompt by saving a new version through `write(operation="prompt_version.save")`.
- Run evaluation experiments
Execute an end-to-end evaluation experiment via Ollie with `run_experiment`, using a test suite, prompt, and scorers.
Tags
- llm
- observability-incident
- data-analysis
- ai-model
- awesome-list
Related MCP servers
Context7 MCP
OfficialOfficial Context7 MCP server that brings up-to-date, version-specific library documentation and code examples into AI coding prompts.
Chrome DevTools MCP
OfficialOfficial Chrome DevTools MCP server for controlling and inspecting a live Chrome browser from coding agents such as Gemini, Claude, Cursor, and Copilot.
Mastra/mcp
OfficialClient implementation for Mastra, providing seamless integration with MCP-compatible AI models and tools.
GitHub MCP Server
OfficialConnect your AI assistant to GitHub repositories, issues, pull requests, and search.