Never run out of usage mid-task

Your coding agents, on your computers, from any screen

Run Claude Code, Codex and more on your own machines, on every plan you have. Follow them from anywhere.

Opening Shuriken on your Mac

  1. Open the downloaded .dmg and drag Shuriken to Applications.
  2. Open Shuriken. macOS says it can't verify it; press Done.
  3. Open System Settings › Privacy & Security, scroll to Security, press Open Anyway next to Shuriken, then confirm.

Why? Shuriken isn't notarised by Apple yet.

Shuriken at work on a project: a workflow, a background agent and a test run in Activity, a chat that moved to another plan when its own ran out, and tasks waiting in the queue

Every source

Every plan you have, side by side

Sign in with your Claude, ChatGPT and Gemini plans, add API keys, and use them together. See what each one has left at a glance.

  • Several plans per provider
  • API keys and custom endpoints
  • Usage and resets for each one
Sources: Claude plans, an API key, ChatGPT and Gemini, each with its usage

Smart switching

Never run out mid-task

When a plan hits its limit, Shuriken moves to the next one by itself and keeps going. Not just the chat: everything that is running moves with it.

What keeps running when a plan hits its limit
Work Other tools Shuriken
The chat Sometimes Yes
Running workflows Stops Yes
Background agents Stops Yes
Commands Stops Yes
A chat's card: Switched to work@example.com, because personal@example.com reached its limit

Built for real work

One account, every device

Works with 67 coding agents

Invite-only for now