Quickstart · BINA Docs

Install BINA, start the API and console, and read a sample bulletin in about five minutes.

By the end of this page you will have the BINA API running on your machine, the console open in a browser, and a bulletin response in your terminal. You do not need any API keys: a fresh checkout serves a built-in sample inventory modelled on the July 2026 bulletin.

#Before you begin

Node.js 20, 22, 24 or 26. The native better-sqlite3 module does not build on Node 18. Check with node -v.
npm, using the committed lockfile.
About 2 GiB free disk space if you later load live data. BINA refuses to fetch when free space falls below that.

#1. Install

Install from the repository root so the native module is built for your Node version:

bash

npm install

#2. Start the API and the console

bash

npm run dev

This starts two processes:

Process	Address
API	`http://127.0.0.1:8787`
Console	`http://127.0.0.1:5173`

If port 5173 is taken, run CLEARY_WEB_PORT=5180 npm run dev instead.

#3. Check that the API is up

bash

curl -s http://127.0.0.1:8787/v1/radar/status | python3 -m json.tool

You should see "mode": "fixture". That means BINA is serving the sample, not live publications.

#4. Fetch a bulletin

Ask for everything published in July 2026:

bash

curl -s "http://127.0.0.1:8787/v1/radar/bulletin?from=2026-07-01&to=2026-07-31" | python3 -m json.tool | head -30

The header object counts what was reviewed, and important lists the level 1 and 2 cases. Each card has a reportPath you can open next.

#5. Open a case report

bash

curl -s http://127.0.0.1:8787/v1/radar/cases/DMA.100075/report | python3 -m json.tool | head -40

Or open the same case in the console: http://127.0.0.1:5173/?case=DMA.100075&month=2026-07.

#Next steps

Follow the tutorial Read your first bulletin to understand what each part of the response means.
Load live decisions to replace the sample with real publications.
Configure hosted extraction to let a model propose parties and summaries.