AnchorRun your first BI dashboard locally (Metabase)

This tutorial helps you launch a local BI stack (Metabase + Postgres) and start exploring data right away. We’ll use the open-source demo project immoteur/webhooks-ingest-demo↗, which includes a small webhook ingestion API and a Metabase setup that reads directly from the database.

If you’d like to explore the result before running anything locally, open the live demo↗.

Tip: Use make demo for live mode (real webhook traffic via smee.io). If you don’t have a source yet, make demo-with-seed injects sample events so your dashboards are not empty.

• A local Metabase instance you can open in your browser.
• A Postgres database populated with webhook events and analytics-friendly tables.
• A first dashboard you can inspect, remix, and share locally.

• Docker + Docker Compose (v2)
• Node.js 20+ (required by the demo scripts)
• Git + Make
• Free ports on your machine (defaults: 8080, 3001, 15432)

• Demo repository: https://github.com/immoteur/webhooks-ingest-demo↗
• Metabase documentation: https://www.metabase.com/docs/latest/↗
• smee.io: https://smee.io/↗

git clone https://github.com/immoteur/webhooks-ingest-demo.git;
cd webhooks-ingest-demo;

Use live mode (real data) with the full Docker Compose stack + smee.io relays:

make demo;

Example output:

[demo] Waiting for API at http://localhost:8080...
[+] Running 1/1
 ✔ Container webhooks_ingest_api                           Healthy   17.6s
 
[demo] Ready (live mode, no seed):
- API:      http://localhost:8080
- Metabase: http://localhost:3001
- Metabase admin: [email protected] / DemoAdmin!2025ChangeMe
- Metabase public dashboard: http://localhost:3001/public/dashboard/16b57a07-9831-4ec9-b927-b6207d3eaa31
 
[demo] smee relays (send webhooks here):
- classified-notification: https://smee.io/classified-notification-3o41g78qykci2otd
- classifieds-export:      https://smee.io/classifieds-export-holes6hi6y85rx31

What this does for you:

• creates .env from .env.example if needed
• starts Postgres + the ingestion API + Metabase
• bootstraps Metabase (admin user + DB connection)
• generates and starts 2 smee.io relays (one per webhook endpoint)
• prints URLs and credentials (including a public Metabase dashboard URL)

Metabase admin credentials also come from .env (defaults in the demo):

• [email protected]
• METABASE_ADMIN_PASSWORD=DemoAdmin!2025ChangeMe

1. Open http://localhost:3001
2. Log in with METABASE_ADMIN_EMAIL / METABASE_ADMIN_PASSWORD
3. Open the demo dashboard (or the public dashboard URL printed by make demo)

In Metabase, you should see tables like:

• webhook_events: raw ingestion log of received webhook requests (with retention cleanup)
• classifieds: flattened records derived from Immoteur payloads (classified-notification / classifieds-export)
• classified_images: one row per image (joined to classifieds)
• classified_price_history: one row per price change (joined to classifieds)

Try a first question in Metabase’s SQL editor:

select count(*) as events from webhook_events;

Real data (Immoteur)

Quick test (cURL)

After make demo, your terminal prints 2 public relay URLs:

TEXT

[demo] smee relays (send webhooks here):
- classified-notification: https://smee.io/...
- classifieds-export:      https://smee.io/...

classifieds-export payloads can be large; if your relay rejects them (for example with HTTP 413), send them directly to the local API (see the cURL tab) or use make demo-with-seed.

Configure your Immoteur webhooks to send events to these smee.io URLs (one per endpoint). When an event is emitted, it’s relayed to your local ingestion API and written to Postgres.

Back in Metabase:

• re-run your question or refresh the dashboard
• check webhook_events count increased

To stop containers (keep data):

make stack-down-smee;

To wipe everything (fresh start, deletes Docker volumes):

make stack-reset-smee;

You now have a local BI stack running: an ingestion API that writes to Postgres, and Metabase reading from it to power dashboards. Next, try creating your own question (e.g., “events per hour”), add it to a dashboard, and point the webhook endpoints to a real source when you’re ready.