Bring your Claude key. The difference is where the agent runs.
TL;DR
This isn't a model comparison. Bring your Anthropic key to rigless and run Claude there too.
Claude Code executes on your local machine. rigless runs every session in an isolated cloud microVM.
rigless works from any browser — no local toolchain, no admin rights required.
If you want Claude running directly on your own machine with your own local tools, Claude Code does that well.
We want to be upfront before this comparison starts: it isn't an argument about which model is smarter. Claude is an excellent model, Anthropic built a well-regarded CLI agent in Claude Code, and if you bring your Anthropic key to rigless, you can run Claude inside rigless too. The interesting question is where the agent actually runs, and what happens to your workflow if you ever want that to change.
Claude Code runs on your own machine — your local filesystem, your local shell, your local environment as the blast radius if something goes wrong. For a lot of developers that's exactly what they want: it's their machine, doing what they tell it to.
rigless takes the opposite default: every session gets its own isolated microVM, a real Ubuntu sandbox that is not your laptop. No local dependency installs polluting your environment, no risk to files outside the sandbox, no need for your laptop to even be on. Reach the same running session from a phone, a tablet, or a different computer entirely, mid-task.
Claude Code is built around Claude. rigless doesn't require that choice: bring an Anthropic key and use Claude, or bring OpenAI, Google, OpenRouter, Moonshot, or Zhipu instead — sometimes in the same afternoon, comparing the same task across models. If you're already committed to Claude, rigless doesn't take that away; it just means you're not locked in if that ever changes.
Claude Code needs a working local dev environment: installed toolchain, configured shell, admin rights on the machine. rigless needs a browser. If you're on a machine you don't fully control, that's a real difference in practice, not just theory.
Claude Code | rigless | |
|---|---|---|
Runs on | Your local machine | Isolated cloud microVM |
Reachable from | The device it's installed on | Any browser, any device |
Model | Claude | Any — including Claude via BYO key |
Setup | Local toolchain + CLI install | None — open a browser |
You want the agent operating directly on your own machine, with your already-configured local tools.
You're fully committed to a local-first workflow.
You want that same caliber of model working in an environment that isn't your laptop.
You want to reach the session from anywhere, not just one machine.
You don't want to be tied to any one model provider if your needs change.