Understand Every Line
~1kB tracker size, 500 lines server code, 10 dbt models. No magic, no abstraction layers. Understand how events flow from browser to BigQuery in 30 minutes.
measureStack The Sovereign Analytics Stack
Full-control, simple, and extensible.
Exact Event Data
Privacy-aware
Scale to Zero
MIT License
Most analytics tools are black boxes. You can't see how events are processed, can't access raw data, can't adapt the schema. measureStack is different: understand the entire codebase, query every event, change anything.
Built for Data Professionals
Query Raw Events
Every event lands in BigQuery as-is. No pre-aggregation, no sampling, no waiting for exports. Be clear what really happens.
Adapt, No Limits
Need a custom field? Add it. Different enrichment logic? Change it. The code is simple enough to modify confidently.
Your Infrastructure
Runs on your GCP project. Data never leaves your environment. Full control over retention, access, and compliance.
Privacy That Makes Sense
Daily-rotating hashes for cookieless tracking. IP truncation. You can explain exactly how it works because you can read it.
dbt-Native Pipeline
Raw events transform via dbt models into session and user tables. Add your own models, join with other data sources, build custom metrics.
Quick Start
Prerequisites: Google Cloud CLI installed and authenticated.
1. Clone & Configure
git clone https://github.com/measurestack/measure-stack.git
cd measure-stack/deploy
cp config.source.template config.source
# Edit config.source with your GCP project ID and domain2. Deploy
./deploy_app.sh # Deploy tracking API
./deploy_dbt.sh # Deploy analytics pipeline3. Add to Your Site
<script src="https://your-tracker-url/measure.js"></script>
<script>
_measure.pageview();
_measure.event('signup_click', { plan: 'pro' });
</script>4. Query Your Data
-- Your events, your SQL, no restrictions
SELECT
DATE(timestamp) as day,
event_name,
COUNT(*) as events,
COUNT(DISTINCT hash) as unique_visitors
FROM `your-project.measure_js.events`
WHERE timestamp > TIMESTAMP_SUB(CURRENT_TIMESTAMP(), INTERVAL 7 DAY)
GROUP BY 1, 2
ORDER BY 1 DESC, 3 DESC;Architecture
Event hits API, gets enriched (device, location), stored in BigQuery. That's it. No queues, no complex pipelines, no hidden processing.
Why Simplicity Matters
When something breaks at 2am, you need to understand what's happening. When stakeholders ask "how is this metric calculated?", you need a real answer. When regulations change, you need to adapt fast.
With ~500 lines of code, you can:
- Debug in minutes - grep the codebase, find the issue, fix it
- Explain your data - show exactly how each field is computed
- Adapt to requirements - add fields, change logic, no vendor tickets
- Onboard teammates - they can read the entire system in a day
Privacy Model
With Consent
- Persistent client ID cookie
- Cross-session tracking
- Return visitor analysis
Without Consent
- Daily-rotating SHA-256 hash
- Same-day sessions only
- No cookies, GDPR-ready
Hash = SHA256(IP + UserAgent + DailySalt). Salt rotates and deletes after 24h. IP truncated before storage. All in src/services/salt.ts - read it yourself.
Comparison
| measure |
Google Analytics | Plausible / Matomo | |
|---|---|---|---|
| Can you read the code? | Yes, ~500 lines | No | Yes, 100k+ lines |
| Raw event SQL access | Direct BigQuery | Export only | Limited / Plugin |
| Modify tracking logic | Edit source | Not possible | Complex plugins |
| Data location | Your GCP project | Google servers | Self-host or SaaS |
| Setup time | minutes | mins (SaaS only) | days (self-host) / mins (SaaS) |
| Pricing | Cloud costs (free tier possible) with scale to zero | Free (you are the product) | from $9+/mo or cloud costs (self-host) |
Stack
Runtime: Bun | Framework: Hono | Storage: BigQuery | Transforms: dbt | Deploy: Cloud Run
All standard tools. If you've used any of these, you already know half the stack.