anfra Core
Run the open-source anfra engine on your own machine or server
anfra Core is the open-source anfra engine, licensed under Apache 2.0. It is a single binary, anfra, that runs on your laptop or on a server. You don't need a cloud account.
How it works
anfra sits between your apps and your semantic layer:
- Your semantic layer is your models, datasets and metrics, defined as code in your project. anfra ships with AMQL built in.
- The anfra SDK runs in each app's page. The page asks for data by dataset and metric names, never SQL, and declares how its filters, drill-downs and clicks connect. See Semantic UI.
- anfra Server (
anfra serve) serves your apps. It turns each request from the SDK into a query on your semantic layer, runs it on your warehouse, and returns the results. It also serves an API and an MCP server, so coding agents can read your models.
Queries run on your warehouse, and results go straight to the page. anfra includes no AI model: you bring your own coding agent, and the anfra skills teach it to model data and build apps.
Ways to run it
| Use it to | |
|---|---|
| The CLI | Build apps on your machine. anfra serve listens on 127.0.0.1:7878, so only you can reach it. |
| Docker | Run a project on a server, or try anfra without installing. Mount the project at /repo: docker run -p 7878:7878 -v "$PWD:/repo" ghcr.io/holistics/anfra. |
Access control
The open-source server has no login or permissions. Anyone who can reach it can open its apps and query the project's data through them.
- On your machine, keep it on the default loopback address.
- On a server, publish the port only to people who may query the data, or put the server behind your own SSO proxy.
Per-viewer permissions (the same app showing different numbers to different users), sharing links and audit logs are part of anfra Cloud, which is not available yet.