Your own always-on, self-hosted cloud lab: real, root-access VMs in the browser, with an AI agent that operates the machine.
Open a real Linux VM from any device: laptop, tablet, or phone. No setup, no waiting.
PequeRoku is an open-source, self-hosted alternative to Replit, Codespaces, and Gitpod. Instead of a locked-down container, you get a real virtual machine (QEMU/KVM) with full root, a browser IDE, and a built-in AI agent that reads your code, runs commands, brings services up, and fixes things inside the box.
It is always on, with no cold starts, no sandbox, and no per-seat billing. The only limits are the ones your own hardware sets.
git clone → docker compose up → http://localhost:8000 → start coding
Every developer knows the ritual:
| Time wasted | 25 to 50 minutes lost on every environment switch, days lost onboarding |
| Dependency hell | version conflicts, missing packages, mismatched OS versions |
| Context switching | cloning repos and rebuilding state for every project |
| "Works on my machine" | environments that drift the moment you look away |
Cloud IDEs promised a fix. In practice they are limited, vendor-locked, and their cost scales with your team. You should not have to choose between power and convenience.
- Real root on a real VM. Install Docker, run
systemd, build a kernel. It is a Linux machine, not a sandbox. - Always on. Your workspace is ready when you are, with no boot and no cold start.
- An agent with real tools. Pequenin edits files, runs your tests, starts servers, and verifies them live inside the VM. See the internals in AI.md.
- Reachable from anywhere. The IDE is responsive, so you can work from a laptop, tablet, or phone.
- Open source and hackable. It runs on your own infrastructure under the MIT license, and every layer is extensible.
- Affordable to self-host. It is built to run on modest hardware, including a homelab.
I built PequeRoku because I could not find a remote dev platform that was open, root-accessible, always-on, reachable from anywhere, and affordable to run myself. So I made one.
Pequenin does not just suggest. It edits, runs, and verifies inside your VM.
Pequenin is an agentic coding agent that operates your environment. It lives in the IDE chat and works through real tools. The full breakdown lives in AI.md:
- Reads, writes, and edits files with whitespace-tolerant, surgical edits.
- Searches your code with glob and grep, directly on the VM.
- Runs shell commands, both foreground and long-running background jobs it can monitor.
- Brings services up and checks that they respond.
- Searches the web and fetches URLs when it needs documentation.
- Spawns subagents to explore in parallel, and keeps multiple switchable conversations per VM.
- OpenAI-compatible, so you bring your own provider: Groq for speed, OpenAI for quality, or HuggingFace for flexibility.
The result is not a snippet to copy. Pequenin creates the files, installs the dependencies, starts the server, and confirms it responds. For the agent loop, tools, and event protocol, read AI.md.
| Capability | Description |
|---|---|
| Real VMs | QEMU/KVM with strong per-developer isolation |
| Browser IDE | Monaco editor and Xterm.js terminal, fully responsive |
| Agentic AI | Pequenin operates the VM: files, shell, services, and web (AI.md) |
| Live preview | Built-in mini-browser proxies your app straight from the VM |
| Persistent and always-on | Your workspace survives, with no cold starts |
| Disposable workspaces | Reset to a clean slate while keeping your config |
| Full root | Install anything, with no guardrails in the way |
| Quotas and roles | Per-user resource limits for team deployments |
Three services and your VM work together:
Browser React IDE (Monaco + Xterm.js)
|
| HTTPS / WebSocket
v
web_service Django, DRF, Channels
| auth, quotas, templates, Pequenin (the AI brain)
| HTTP
v
vm_service FastAPI
| VM lifecycle (QEMU), SSH pool, live terminal
| SSH / SFTP
v
Your VM Debian, full root, /app workspace
The brain (the AI and agentic loop, detailed in AI.md) runs in web_service. The hands (files, exec, terminal) are SSH operations that vm_service executes inside the VM. The front end only ever talks to web_service.
Stack: Django · DRF · Channels · FastAPI · QEMU/KVM · React · Monaco · Xterm.js · Postgres · Redis · Docker
Full architecture and setup: Wiki. Deep dive on the AI engine: AI.md.
git clone https://github.com/HectorPulido/pequeroku.git
cd pequeroku/source
docker compose up # Docker Compose supported out of the box- Open
http://localhost:8000. - Log in and create your first VM.
- Start coding, or ask Pequenin to scaffold the project for you.
Detailed walkthrough: Wiki, Getting Started.
That gives you a self-hosted, Replit-style workspace under your control.
| PequeRoku | Replit / Codespaces / Gitpod | |
|---|---|---|
| Hosting | Self-hosted, your hardware | Their cloud |
| Compute | Real VM (QEMU/KVM) | Containers |
| Root access | Full root | Limited |
| Cold starts | None, always-on | Boot and wait |
| AI | Agent that operates the box | Mostly autocomplete or chat |
| Source | Open source (MIT) | Proprietary |
| Cost | Your hardware | Per-seat or usage |
- Fast snapshots and one-click rollback
- Richer multi-user roles and guardrails
- Automations: a push triggers tests in the active VM
- Better UI for managing multiple instances
- gRPC internal transport and a standalone AI microservice (see the roadmap in AI.md)
- More Pequenin capabilities
PequeRoku is open to ideas, bug reports, and pull requests.
- Browse the Issues.
- Share feedback, ideas, or problems. I reply to everything.
- Deploy it in your homelab and tell me how it went.
If PequeRoku resonates with you:
- Star the repo. It helps others discover it.
- Spread the word.
- Run it in your homelab and share your setup.
PequeRoku is not the ultimate platform. It is your platform, a small way to take back control.
- pequeroku.net
- GitHub Wiki
- Medium Article
- License: MIT. See the LICENSE file for details.
 Â