Skip to content

Uncage

Your website gives the instructions. Your user’s own Claude Code does the work, on their computer, and calls your website’s tools. The process and the results stay on your site.

Uncage is a set of npm packages that put Claude Code (and Codex) behind your own web UI, with approvals, streaming, Stop and reload-safe history.

It has two modes. Both use the same protocol, the same React hooks and the same <AgentChat> UI.

Connected Mode (new in 0.2) Local Mode
Who runs the agent Each end user, on their own computer, with their own Claude Code login You, on the machine that runs your app
What you add to your site @uncage/hub in your Node backend, <AgentChat> in a page, an approval page at /device A same-origin proxy route and <AgentChat>
What the user does Runs one command once and approves the computer in the browser: curl -fsSL https://your.site/api/uncage/install.sh | bash on macOS and Linux, irm https://your.site/api/uncage/install.ps1 | iex on Windows Nothing: it is your machine
Good for SaaS products where users have Claude Code and want it to work with your product Internal tools, desktop-style apps, developer dashboards
  • A chat on your website that drives their Claude Code. Nothing runs on your servers except your own code.
  • Live streaming of the answer and of every tool step.
  • Allow / Deny cards for commands, file edits and any of your tools that need confirmation.
  • Stop, attachments, model, effort and permission-mode pickers, and history that survives a reload.
  • With site storage on, the conversation stays readable on your site even when their computer is off.
  • React Native apps use @uncage/client and @uncage/react unchanged: streaming, approvals, photos, resume after background, approve a computer from a phone.
  • Browser extensions run <AgentChat> in a side panel against your hub, with trustedOrigins.
  • One agent, every site: a user installs the agent once, and every site that embeds Uncage connects in one click.
  • Site tools. Declare a capability with defineTool (a name, a zod schema, a handler). Claude Code sees it as an MCP tool. The handler runs in your backend, as the signed-in user, and can pull files from the run’s workspace (ctx.files.pull).
  • Skills. Give Claude Code your house rules and templates as plain markdown. It loads them like any Claude Code skill.
  • Per-run choice. selectManifest picks the tools and skills for each run from the prompt and the user.
  • Your auth, your storage, your scale. getUser(req) plugs in your sign-in. Postgres and Redis adapters let the hub run on several nodes behind a load balancer.
  • It never sends code or shell commands from your website to a user’s machine. The hub sends only data: tool names and schemas, skill text, and a fixed set of commands such as “start a run”.
  • It never shares one Claude login among users. Every machine uses its owner’s own Claude Code.
  • A website can never switch on “bypass permissions”. Only the user can, in a file on their own computer.

See the trust model for the full list.