Skip to main content
Signing in with resumecontext auth binds that machine. Every session it records carries that machine’s identity, and every result your agent gets back carries an origin:

​
Why it matters

History is not proof that anything it describes exists here. A session from another machine can reference files, checkouts, datasets, running services, environment variables or credentials that are simply absent locally. So when your agent reuses a command or path from a session that is not this_machine, it is instructed to say where it came from and check the inputs exist before running it.
origin is the only evidence of which machine. Paths inside a transcript are just text that session recorded. Teammates commonly share a directory layout, and /Users/<name>/work/repo can exist on several machines or none. A familiar-looking path does not mean the work happened here.

​
Adding another of your machines

To sync a project from a second machine of your own (a desktop, a laptop, a server), you don’t need an invite or accept. You already have access, so the machine only needs to sign in and be pointed at the project folder.
1

Install the CLI

2

Sign in with the same account

This binds the new machine, so its sessions are recorded as coming from it, and starts the background sync service.
3

Go to the project folder

The folder must contain the project’s .resumecontext.json, usually because it is committed and you cloned the repository. If it isn’t committed, copy the file from your other machine into the same folder.
4

Add the folder to this machine's auto-sync

Because .resumecontext.json is already there, init connects this folder to the existing project instead of creating a new one. It checks your access, asks which coding agents to sync from on this machine, and adds the folder to auto-sync. The file itself is left unchanged.
Run init in the folder that contains .resumecontext.json, not in a subfolder. In a subfolder, init creates a separate new project.
resumecontext daemon start is not enough on its own. It only makes sure the background service is running, and it does not add the folder you run it in. A folder is added by init, or by any project command run inside it, such as sync or mcp.
Confirm the folder is registered and auto-sync is running:
Sessions from the new machine show up with origin set to same_person_other_machine when your agent searches from your other machines.

​
The web app

The dashboard shows origin per session too, but a browser is not a bound machine, so your own sessions show as unknown there rather than claiming to be this_machine.