Deploy your app
in minutes,
not weekends.
Add a server, connect a Git repo, click Deploy. Code</>sensei handles everything between your code and a live URL — so you don't have to.
Live now · start free, upgrade any time.
$ sensei deploy --app shipfast
→ Connecting to server…
✓ Server ready
✓ Code pulled @ a7c4f2e
✓ Dependencies installed
✓ Database migrated
✓ Static assets collected
✓ App server restarted
✓ Web server reloaded
Deploy complete in 38s.
URL: https://shipfast.codesensei.cloud
→
Three steps from zero to deployed
Bring your server, point us at your repo, hit Deploy. The platform handles the rest.
Add a server
Drop in an IP and SSH credentials. We detect the OS and link it to your account.
Create a project
Point at your Git repo, pick a branch, set environment variables.
Deploy
Hit Deploy. The platform handles the rest and streams the log straight to your browser.
Simple, transparent pricing
Starter at $5/mo. Scale as you grow.
Starter
For solo projects and personal apps.
- 1 server
- 2 projects
- Unlimited deploys
- Community support
Pro
For founders shipping in production.
- 5 servers
- 10 projects
- Email support
- Deploy webhooks
Team
For larger workloads.
- Unlimited servers
- Unlimited projects
- Unlimited deploys
- Priority support
All plans include unlimited deploys.
A deployment platform that respects your stack
Code</>sensei is a web-based deployment platform for Django and Node.js applications. It takes the work of getting your code from a Git repository to a running, production-ready service on a Linux server you own — provisioning the operating-system packages, configuring the application process, wiring up the reverse proxy, terminating TLS, running database migrations, and serving the deploy log back to your browser in real time — and reduces it to one button click. There is no command-line tool to install on your laptop, no YAML pipeline to maintain in your repository, and no agent process to keep updated on your server.
The deploy targets are servers you control. A Hetzner Cloud instance, a DigitalOcean droplet, a Linode, an AWS EC2 box, a Proxmox VM in your garage — anything that runs Debian 12, Ubuntu 22.04 / 24.04 LTS, AlmaLinux 9, Rocky Linux 9, or FreeBSD 14 and that you can give us an SSH login for. You bring the server; we bring the platform that knows how to make your code run on it. Your application binaries, your application data, and the database backing it never leave the box you added. If you cancel your subscription tomorrow, what we deployed keeps running unchanged — no vendor lock-in.
Built for the two stacks most teams ship
First-class support for Django and Node.js — including the frameworks that build on top of them.
Django and Python
Standard requirements.txt
+ manage.py layouts deploy
out of the box. The platform creates a virtual environment, installs your dependencies,
runs migrate and
collectstatic, writes a
systemd unit that runs gunicorn (WSGI) or Daphne (ASGI / Channels), and serves your
static assets directly from the edge with a one-year cache. Works with Django Rest
Framework, Django Channels, Celery, and the rest of the standard ecosystem without
any extra configuration on your part.
Generic Python projects deploy the same way. FastAPI, Flask, Starlette, custom WSGI
apps — anything with a requirements.txt
and a documented entrypoint can ship.
Node.js and JavaScript
Reproducible installs via npm ci
from your lockfile, an automatic build step when your
package.json defines a
build script, and a
systemd unit that runs npm start
with automatic restart on crash. Next.js, NestJS, Express, Fastify, Astro server mode —
all supported with no per-framework configuration.
Static builds (Astro static, Vite, plain HTML) skip the systemd step entirely. The edge
proxy serves your dist/
directory directly with one-year cache headers, dropping runtime memory to zero.
All the deploy boilerplate, handled
The infrastructure pieces every production app needs — pre-configured on every server you add.
Reverse proxy with TLS
Caddy at the edge with automatic Let's Encrypt certificates. HTTPS the first time the deployed URL is hit, automatic renewal forever after. Custom domains via a single CNAME and one click in the dashboard.
systemd-managed processes
Each project runs as its own service unit with the correct working directory, environment file, restart policy, and resource limits. Survives reboots, restarts on crash, logs go to journald.
Release directory layout
Capistrano-style timestamped releases with an atomic
current symlink. The
previous five releases stay on disk so a roll-back is a single symlink flip — no
rebuild, no re-install.
Streamed deploy logs
Every command the platform runs on your server is streamed to your browser through a WebSocket and persisted on the deployment row. No black box — you can read exactly what ran and when.
Environment variables
One JSON blob on the project form becomes a shared
.env file on the server,
referenced from the systemd unit so build-time and run-time see the same values.
Surviving across releases is the default.
SSH credentials at rest
Server credentials sit Fernet-encrypted on the platform database. The decryption key lives in process memory only, never in the database or any image artifact. Read the security model for the full picture.
Designed around what hurts
Your data stays on your server.
Code</>sensei is not a hosting provider. Your application binaries, your database, and your user-uploaded files all live on the Linux server you added. The platform opens an SSH session when there is work to do — a deploy, a roll-back, a bootstrap — and closes it when the work is finished. There is no agent process polling the control plane on your behalf. No persistent inbound connection to your server when you are not actively deploying.
Zero vendor lock-in.
What the platform deploys is a Capistrano-style release directory and a systemd unit file. If you ever want to walk away, the files on your server keep running unchanged and you can take over the deploy loop with a one-line shell script. We give you a path out by design.
Roll back in seconds.
The previous five successful releases stay on disk after each deploy. Rolling back to any of them is a symlink flip and a systemd restart — no rebuild, no re-install, no re-run of migrations. Click the deployment you want to restore, confirm, and your previous version is live again before the next coffee sip.
Pricing that scales with usage, not with seats.
$5/month gets you one server and two projects. $19/month gets you five servers and ten projects. $79/month removes the limits. We charge for the platform, not for the number of human heads on your team. Every plan includes unlimited deploys — there is no "build minutes" cap or per-deploy fee. See the full pricing page for the breakdown.
Ship something before lunch.
Drop in a server, point us at your repo. We'll handle the rest.
Add your first serverOr read the 10-minute quickstart guide first.