Portal
The web interface for organizing workspaces and projects, browsing runs, visualizing metrics, and comparing experiments.
Metrana Docs
Track, visualize, and compare your machine learning experiments.
Welcome to the Metrana documentation. Documentation is organized by product.
Python SDK
The metrana client library your training code uses to log metrics,
attributes, and reinforcement-learning data — with durable, fail-loud delivery.
Ingestion API
Create runs and log attributes, metrics, reinforcement-learning data and agentic traces from any language, over gRPC or REST.
Query API
Read runs, attributes, series and traces back out — filter, summarize, bucket for charts and aggregate across environments.
MCP Server
Connect your own agent — Claude Code, Cursor and other MCP clients — to read runs, metrics and RL environments across your organization’s workspaces.
RL guide
Using Metrana for reinforcement learning? A quickstart of links through the RL features — logging, querying, and viewing environments and episodes.
More tools coming
Documentation for additional Metrana tools (CLI and more) will live here as it’s published.