Infrastructure intelligence and trusted metrics, together at launch.
Everything data engineers and business leaders need in one open-source workspace. Deep stack understanding, canonical KPIs, and Company Brain reasoning, no waiting for the next release.
The technical foundation for data teams
Context-aware editors wired to your live catalog, a semantic graph that stays fresh from metadata crawls, and unified infrastructure understanding across orchestration and warehouse layers.
Self-host with Docker in minutes: metadata-only, credentials never leave your network.
- Stack-aware SQL, dbt, and Python editors with inline lineage
- Living catalog from Snowflake, dbt, Airflow, and 40+ connectors
- Cross-layer impact analysis in seconds, not Slack threads
- Apache 2.0 · RBAC · audit logs · VPC-ready deployment
Business knowledge, certified metrics, a graph agents can query
Define revenue, churn, and activation once with full lineage, then give agents the definitions, policies, and relationships they need to answer accurately.
Headless, lightweight MCP for Claude Code, Codex, Cursor, and custom agents. Reconciliation so finance and product agree on the same number.
- Business knowledge: definitions, caveats, policies, owners
- Agent query engine: ask, plan grain, compile governed SQL
- One source of truth per KPI with certification workflow
- Same certified context across BI, notebooks, and MCP agents
General-purpose agent plus a custom builder
A ready agent with dbt, Dagster, data analysis, text-to-SQL, RAG, and CSV — or build your own with the tools you choose.
Every agent is grounded in the Context Layer, your Company Brain, then connects to Slack, Gmail, GitHub, and the rest of the ecosystem.
- Specialized categories for the work data teams already do
- Attach web search, Slack, Gmail, GitHub, and more
- Workflows: weekly analysis to Slack, PDFs to email
- Metroflow’s outcome, working as agents in your stack