session_resume

Resumes a stopped session on a fresh machine, with the same id and everything as it was when it stopped: tabs and their history, logins, form values and the files in /workspace. The tools that act on a session do this on their own when you give them the id of a stopped session; the read-only ones do not, so call this first for them.

Inputs

sessionId is the id that session_create returned. The server keeps no default session. This tool works on the session as it is, running or stopped, and does not resume it first.

Example call

This is what an MCP client sends. You do not write it yourself: ask the model in plain words and it calls the tool.

{  "name": "session_resume",  "arguments": {    "sessionId": "6c1f3c9e-1d2a-4f67-9b0e-1b2c3d4e5f60"  }}

What it returns

The session’s id, its live view URL (stdio only, and null for a session without a browser) and the workspace path, as JSON text.

{  "sessionId": "6c1f3c9e-1d2a-4f67-9b0e-1b2c3d4e5f60",  "liveUrl": "https://…/live/6c1f3c9e-…?token=…",  "workspace": "/workspace"}

When a call fails, the result is marked as an error. Its text starts with the API’s error code, such as unauthorized or session_not_running, followed by its message.

Annotations

The server tells every client what this tool does to the world, so a client can decide what to ask you about before it runs.

Good to know

• A stopped session that saved nothing (its machine was lost before its first checkpoint) cannot be resumed: the call fails with nothing_saved.
• Running processes and the browser’s caches do not come back, and pages reload.
• Resuming counts towards your plan’s concurrent sessions and is refused when a spending limit is reached.
• The live view URL lets anyone who has it control the browser. Share it only with people who should.

The same thing with the API