Deploy and Host HitKeep

Deploy HitKeep, an open-source privacy-first analytics platform you can self-host as one Railway service.

This template uses the official pascalebeier/hitkeep:2.7.0 Docker image, exposes port 8080, attaches a persistent Railway volume at /var/lib/hitkeep/data for active DuckDB data, and provisions a Railway Bucket for backup/archive Parquet snapshots.

About Hosting HitKeep

HitKeep is a single-binary web analytics app with embedded DuckDB and embedded NSQ. It tracks pageviews, events, goals, funnels, ecommerce activity, AI crawler visibility, Search Console aggregates, exports, and read-only API/MCP access without requiring PostgreSQL, Redis, ClickHouse, Kafka, or a separate queue.

On Railway, the app runs as one persistent service backed by one attached volume and one private S3-compatible Railway Bucket. The template generates a unique HITKEEP_JWT_SECRET during deployment and configures HitKeep to use the generated Railway public domain.

Why Deploy HitKeep on Railway

One service instead of a multi-container analytics stack
Persistent DuckDB storage through Railway Volumes
Off-volume backup/archive snapshots through Railway Buckets
Public HTTPS domain generated by Railway
Per-deployment generated JWT secret
Fast pilot path for self-hosted analytics
Easy migration path to a custom domain later

Common Use Cases

Privacy-first web analytics for small teams
Self-hosted alternative to GA4-style reporting
Cookieless pageview and event tracking
Conversion dashboards with goals and funnels
Search Console aggregate reporting
AI crawler and AI-referred traffic visibility
Internal reporting through exports and API clients

Dependencies for HitKeep

HitKeep does not require an external database, cache, or queue. The required runtime dependency is persistent filesystem storage for its DuckDB data directory.

Deployment Dependencies

Dependency	Railway resource	Purpose
HitKeep runtime	Service from `pascalebeier/hitkeep:2.7.0`	Runs the app, dashboard, tracker, embedded queue, and workers
Persistent data	Railway Volume at `/var/lib/hitkeep/data`	Stores active `hitkeep.db` and tenant databases
Backup/archive data	Railway Bucket named `hitkeep-backups`	Stores HitKeep backup and retention archive snapshots as Parquet
Public URL	Railway service domain	Serves the dashboard and tracker script
JWT secret	Template variable generator	Creates a unique session/signing secret for each deployment

After Deployment

Open the generated Railway domain.
Create the initial admin user.
Add your first site.
Install the tracker by loading https://your-hitkeep-domain/hk.js from your site.

If you later attach a custom domain, update HITKEEP_PUBLIC_URL to the final HTTPS origin so tracker URLs, email links, and redirects stay correct.

Operational Notes

Run one replica by default. HitKeep uses embedded DuckDB and NSQ, so a single-service deployment is the intended low-ops path.
Keep the volume attached. Without it, analytics data will be ephemeral.
The template writes HitKeep automatic backups and retention archives to the Railway Bucket, not the local volume.
HITKEEP_BACKUP_RETENTION only prunes local filesystem backups in HitKeep 2.7.0. Railway Buckets currently do not provide bucket lifecycle configuration, so add a cleanup job if you need bounded object storage growth.
SMTP-dependent features require Railway Pro or above because Railway disables outbound SMTP on Free, Trial, and Hobby plans.

Deploy hitkeep-bucket-template