از Git، از Dockerfile، از ایمیج آماده یا از سایت استاتیک مستقر کنید. Harbora میسازدش، با SSL خودکار دامنه میدهد، سلامتش را میسنجد و تنها وقتی ترافیک را عوض میکند که نسخهٔ تازه واقعاً بالا آمده باشد.
$ harbora deploy
→ checkout main@a1b2c3d
→ build detected Node 22
→ push harbora/shop:build-42
→ deploy starting alongside build-41
→ health GET /healthz 200 OK
→ cutover traffic switched
→ retire build-41 removed
✓ live https://shop.example.com
rollback anytime: harbora rollback 41یک پوستهٔ نازک روی docker run نیست — همان بخشهایی که درستکردنشان خستهکننده است، همانهایی هستند که Harbora خودش برعهده میگیرد.
Git repository, Dockerfile, prebuilt image, static site or a one-click template. The stack is auto-detected when there's no Dockerfile.
Every app gets a domain and a Let's Encrypt certificate, renewed automatically. HTTP redirects to HTTPS by default.
The new container starts alongside the old one. Traffic switches only after health checks pass, so a failed deploy never takes the site down.
Rollback re-releases the previous image instead of rebuilding — you get the exact bytes that were working, in seconds.
Provision PostgreSQL, MySQL, Redis and more in one click. Scheduled backups are encrypted, checksummed and verifiable before you restore.
Live build and runtime logs, CPU and memory per app, and alerts to email, Telegram or a webhook when something breaks.
از یک سرور خالی تا اپی که اجرا میشود و گواهی معتبر دارد.
Link a Git repository or point at an image. Harbora detects the stack and prepares the build.
Choose CPU and memory, then use the free subdomain you get or bring your own.
Watch the build stream live. When health checks pass, traffic switches over — and stays on the old version if they don't.
سهمیهها را خود پلتفرم اعمال میکند، نه اعتماد.
No. If your repository has no Dockerfile, the stack is detected and a production image is generated for you. If you do have one, it's used as-is.
Nothing visible to your users. The new container is removed and the previous version keeps serving traffic — it is never stopped before the replacement is healthy.
Yes. Everything runs as standard Docker containers behind Traefik on your own server, and your images and volumes stay on that machine.
Each workspace gets its own Docker network, and every query is scoped to the workspace at the database level, not just in the UI.
Archives are encrypted before they leave the server and checksummed. A restore verifies the archive first and refuses to run if it doesn't match.