metaxisdata

Language / 语言: English | 中文

Metaxisdata

Metaxisdata is a self-hosted data governance and metadata platform. It connects to MySQL/TiDB and PostgreSQL instances, syncs their schemas into a searchable metadata registry, and derives table- and column-level lineage from SQL definitions and ingested OpenLineage events.

Features

• Metadata registry — Sync the schemas of your instances into one searchable registry. Every table and column carries its DDL and a history of what changed.
• Table and column lineage — Lineage is analyzed from view, materialized-view, and manual SQL, precise down to the column. Expand the graph by depth, or follow a single column's trail.
• OpenLineage — Ingest run events, map namespaces, and issue API keys. Browse jobs, datasets, and events, with links back to Airflow.
• SQL explanation — LLM-assisted explanation of a SQL statement, scoped to the metadata you select, with caching.
• Access control and audit — Users, groups, named roles, and an access-policy editor. Audited operations are recorded in an audit log that is kept forever.
• Agent integration — The mxd CLI (cli/README.md) gives an agent one JSON document per command, stable exit codes, and device login. With MCP enabled, the registry is also available as read-only tools behind an OAuth 2.1 authorization server (docs/mcp.md).

Quick start

You need Docker and a PostgreSQL database the container can reach. Images are published to GHCR for linux/amd64 and linux/arm64:

docker run -d --name metaxisdata \
  -p 8083:8083 \
  -e PG_URL='postgres://user:password@db-host:5432/metaxisdata?sslmode=disable' \
  ghcr.io/ranxy/metaxisdata:latest

Without Docker at hand, every release also publishes the same server as static binaries for five platforms — download one from GitHub Releases and run it directly; the deploy guide's Get the server section shows how.

Open http://<host>:8083 and:

1. Create the first account. It becomes the workspace administrator.
2. Add an instance and sync its schema.

Optional next steps:

• Close self-service signup. Turn on Disallow self-service signup in Settings → General, so only administrators can create users.
• Serve HTTPS. The container serves plain HTTP by design. Terminate HTTPS in a reverse proxy in front of it, and set METAXISDATA_TRUSTED_PROXIES to the proxy's address so the audit log records clients instead of the proxy.
• Protect stored credentials. Set METAXISDATA_ENCRYPTION_KEY so a database dump alone cannot decrypt the credentials stored for your instances — and keep that key safe: losing it leaves the stored credentials unreadable.
• Connect MCP clients. Enable the MCP endpoint in Settings → General — it requires the workspace external URL — then follow docs/mcp.md.

The complete deployment guide — every environment variable, the health and version endpoints, the prebuilt binary, and building the image or the binary yourself — is docs/deploy.md.

Trying it locally

docker-compose.yml starts PostgreSQL and the server together, so one command is enough to look at a built image on your own machine:

make docker-up      # build the image, then start PostgreSQL + the server on http://localhost:8083
make docker-down    # stop both; the compose data volume is kept

Open http://localhost:8083 and create the first account. The stack makes taking a build for a spin easy — published port, hardcoded development password — and is not meant for production. Details and caveats are in docs/local-trial.md.

Screenshots

Metadata browser — the table list of a synced database, with row counts, sizes, columns, and indexes.

Table and column lineage — a table's upstream and downstream edges, expanded down to per-column trails.

How it works

• One Go server binary serves the ConnectRPC API and the frontend — a Vue 3 single-page application embedded in the image — on port 8083. When MCP is enabled, the MCP endpoint is served on the same port.
• PostgreSQL is the only external dependency. The schema migrates itself on startup, so a fresh database needs nothing but a connection URL.
• Background runners sync schemas, analyze and validate lineage, and run maintenance tasks.

Development

# Backend — requires PG_URL; port 8083 matches the Vite proxy
PG_URL='postgres://dev:dev@localhost:5432/metaxisdata?sslmode=disable' go run ./backend/bin/server/main.go --debug

# Frontend — Node 24+, Vite on :3000, proxying API, OAuth, and MCP routes to the backend
pnpm --dir frontend install
pnpm --dir frontend dev

# Build
make build           # dev profile
make build-release   # release profile
make build-embed     # release profile with the SPA embedded in the binary
make build-cli       # the mxd client, ./build/mxd

A build without the embed tag serves a placeholder page instead of the SPA: run the Vite dev server, or build with make build-embed.

go test ./... needs no database. The integration suites run the real server against PostgreSQL and MySQL behind the integration build tag, and skip when Docker is unavailable:

make test-integration-smoke
make test-integration

Development conventions, lint rules, and the full command set: AGENTS.md.

Tech stack

• Backend: Go, PostgreSQL, ConnectRPC (gRPC/HTTP), Protobuf / buf
• Frontend: Vue 3, TypeScript, Vite, Tailwind CSS, shadcn-vue
• CLI: Go