// Build together · Ship it live · Prove it serves
One sovereign platform where your developers and your AI agents build side by side, on compute you own. Ship what they make to the world — then watch the numbers prove it’s working. Your code never leaves your servers.
Hand a task to an agent. It spawns an isolated sandbox on your own compute, reproduces the bug, patches it, runs the suite, and opens a PR — without ever touching your machine.
// sandbox · worker-7-9f2a
The same primitive for both. Your developers work in the browser or over SSH; your AI agents work through the API — on identical, isolated sandboxes, spawned in seconds. One team, one workspace, building the same thing together.
Margot
Design
Reviews a PR from her iPad — no laptop, no setup.
Hugo
Backend
Spins up a Postgres-backed sandbox in seconds.
Atlas
Coding agent
Fixes flaky tests across repos overnight.
Inès
Data
Runs a heavy notebook on a beefy node.
Tomas
SRE
Debugs prod in a throwaway replica.
Nova
Migration agent
Runs the nightly DB migration on its own.
Lena
Frontend
Lives in code-server, straight from the browser.
Pixel
Review agent
Combs the diff before you even wake up.
Spunto is European software that runs on compute you own. No host ports, no hidden routing, no data leaving your network — verifiable, not just promised.
Your data never
leaves your servers.
Bring your own compute: workers and services run on isolated Docker bridge networks with no host ports published. Pick the region — or your own rack.
Median spawn time
Secrets encrypted, injected as env vars
Isolated Docker network per worker
Devcontainer-spec compatible
Outbound-only agent — no inbound firewall changes
One-time SSH tokens · 5-minute TTL
Versioned, pre-built environment images
Zero vendor lock-in — it's your stack
Schedule workers and services onto your own nodes — in the EU region of your choice, on-prem, or inside your VPC. Data never crosses a border you didn't pick.
Forget "it works on my machine". People and agents alike get an identical, isolated Docker workspace — pre-loaded with repos, secrets, extensions, and dotfiles. Spawned in under 15 seconds, so they can start building side by side.
api-gateway
node:20 · 3 repos
Smart node scheduling
Workers are placed on the least-loaded online node in your fleet automatically. No manual targeting.
VS Code in the browser
code-server comes pre-installed. Full VS Code experience from any device — no local setup, ever.
Repos cloned automatically
Declare your repositories in the project config. SSH deploy keys are generated and injected at spawn.
Secrets & dotfiles injected
AES-256 encrypted secrets become env vars in every worker. Your dotfiles repo is cloned and applied.
AI-agent ready
Spawn isolated sandboxes via API for your coding agents. Full terminal access, concurrent-safe.
SSH access
Generate a one-time SSH token from the dashboard (5-min TTL). Connect with your regular SSH client — no VPN, no bastion host.
Secure port forwarding
Any port your container listens on is instantly accessible at worker-{n}-{id}-{port}.domain — authenticated, no config needed.
The moment it's ready, it ships to the world. Run Docker services in production on the same nodes as your sandboxes — custom domains, automatic TLS, pre-deploy hooks — without the complexity of Kubernetes.
api
my-app:latest
postgres
postgres:16
redis
redis:7-alpine
pre-deploy hook — api
$ npx prisma migrate deploy
✓ Applied 2 migrations
Deploying new container…
Pre-deploy hooks
Run database migrations, cache warmups, or any command before your container goes live. If it fails, the deploy is aborted automatically.
Custom domains & auto TLS
Point your CNAME at Spunto. Verify ownership with a DNS TXT record. Traefik provisions a Let's Encrypt certificate automatically.
Force-pull & zero-downtime
Deploy pulls the latest image, runs hooks, stops the old container, and starts the new one. Version history lets you roll back instantly.
Shared private network
All services in a deployment share a private Docker network. They talk to each other by service name — no IPs, no service discovery config.
Job runner built in
Define recurring jobs per service — one-off migrations, cron scripts, data exports. Run on demand with full log history.
Shipping is only half the story. Watch what you built reach real people — live logs, resource stats, and usage in one place — so you can prove it's actually working, then keep making it better.
people served
activation rate
satisfaction
support load
daily active · trending up
people it reached
1,240 people used what your team and its agents shipped this month — running on compute you own, in your region.
Available now
Live log streaming
Tail your container stdout/stderr in real-time from the dashboard or via SSH.
CPU & memory stats
Live resource snapshot per worker and per deployment service. Updated every few seconds.
On-demand port access
Any port your container listens on is accessible at worker-X-{id}-{port}.domain — no config needed.
Job run history
Every job execution stores its full log output and exit status. Searchable, paginated, permanent.
On the roadmap
847
Req/s
48ms
p99 latency
0.02%
Error rate
Distributed traces
Connect any server, VM, or cloud instance as a compute node. Your workloads run where you decide — no lock-in, no hidden routing.
In the dashboard, add a compute node to your organization. You get a unique token — that's all you need.
Deploy the agent on your server with a single Docker command. It connects outbound — no firewall rules, no open ports.
The node appears online in seconds. Workspaces, deployments, and production services all schedule on your compute automatically.
Stop juggling five tools for dev environments, deployments, and monitoring. Spunto is the single platform your team actually needs — and it runs on your servers.