OpenCode crossed 100,000 GitHub stars and topped Hacker News because it does one thing the subscription CLIs won't: it talks to any model and never takes a cut of your tokens. Same Claude or GPT, no markup. The catch is the same as every terminal agent — it runs bash on whatever machine you put it on. Here's how to get the best of it without the risk: run OpenCode on its own always-on Firecracker microVM. Any provider you want, a persistent workspace, hardware isolation from your laptop — booting in about 124ms.
What OpenCode is
OpenCode is an open-source (MIT) coding agent for the terminal, from the team behind SST. It opens a rich TUI, connects to the provider you choose, and writes, debugs, refactors, and documents code from the command line. It's provider-agnostic by design — 75+ providers, including fully local models via Ollama — with LSP-powered, type-aware edits and parallel multi-session support.
The pitch is control: you own the model choice, you own the keys, and the core agent is free. The thing it needs is a place to run with a shell and your files — always on, holding your project.
Why a real microVM for OpenCode
A coding agent runs bash. It installs packages, edits files, and executes whatever the model decides — code no human reviewed. On your laptop that's your whole machine on the line. A shared container isn't the answer either: it shares the host kernel, with 200-plus known weaknesses in the stack.
jurniti gives OpenCode its own Firecracker microVM — its own kernel, a KVM hardware boundary, one tenant per box. The same isolation serverless platforms use. Give the agent root inside its own VM and a bad command — or a poisoned dependency, or a prompt-injection attack — stays contained to that box. Reprovision and you're clean in seconds.
BYOK is where OpenCode shines, and jurniti makes it architectural. If you already pay for Anthropic or OpenAI API credits, OpenCode with your own key is the cheapest way to a Claude- or GPT-powered terminal agent — no subscription, no markup. Run opencode auth login, point it at any provider, and the key lives only in your VM. Traffic leaves straight for the provider. jurniti is never in the path and never marks up a token.
OpenCode runs comfortably on Starter (1 vCPU / 2 GiB). And it's reversible: a 30-day money-back guarantee, workspace kept 7 days after cancellation.
Run OpenCode in 3 steps
By hand this is a VPS, the install script, provider config, a systemd unit, tmux so sessions survive a disconnect, and patching for as long as you run it. jurniti runs the box; you keep the agent.
1. Pick a plan and pay
Starter is enough for OpenCode. Check out, and the provision chain runs automatically the moment payment confirms.
2. Your microVM boots with OpenCode ready
A fresh Ubuntu microVM boots with OpenCode pre-installed and on your PATH. You get an in-browser terminal — nothing to SSH into. Configure a provider and go:
opencode auth login # pick any provider, paste your key
opencode
3. Put it to work
Drop into a project, hand OpenCode a task, and walk away. Your config and custom agents live on the persistent home, so they survive every restart.
What people run OpenCode for
- Model freedom. Swap between Claude, GPT, Gemini, Groq, or a local Ollama model without changing tools.
- Cheapest Claude-in-the-terminal. BYOK against API credits you already have, no subscription on top.
- Long, thorough runs. OpenCode trades a little speed for depth and real LSP diagnostics — let it grind on a box that's always on.
- A pinned environment. Your config, your custom agents, your provider keys — set up once, always there.
Snapshot a dialed-in workspace into a template and fork it for the next project — your setup, never your credentials.
jurniti vs a laptop vs a DIY VPS
| jurniti | Your laptop | DIY VPS | |
|---|---|---|---|
| Always on | Yes | No — it sleeps | Yes |
Isolation for bash | Firecracker microVM (KVM) | Your whole machine | Shared kernel |
| Provider keys | Yours, never leave the VM | Yours | Yours |
| Token markup | None (true BYOK) | n/a | None |
| Setup | Automatic · 124ms boot | Local install | Manual |
| Refund | 30-day money-back | n/a | None |
Pricing
Starter is $25/mo ($250/year, two months free) and comfortably runs OpenCode. Pro is $49/mo and Max $99/mo for heavier workspaces or a custom TLS subdomain. Every plan includes the 30-day money-back guarantee, and your model spend is separate and never marked up.
Running a different agent? See the guides for Claude Code, OpenAI Codex CLI, and Goose, or compare plans on the pricing page.