Run a Hub on Your Machine
The lightest way to reach Treeline from other devices and AI agents is to serve from a machine you already use. Your laptop or home desktop becomes the Hub, serving its own Treeline data — nothing to deploy.
Start serving
Section titled “Start serving”tl serveOn first run this creates a master token at ~/.treeline/hub-token and prints it. The token is only used to authorize new devices and agents at the /authorize page — you don’t paste it anywhere else. (Set TL_HUB_TOKEN if you’d rather pin a specific value.)
By default tl serve binds to 127.0.0.1:4242, reachable only from this machine. To reach it from other devices on your network, bind to all interfaces:
tl serve --host 0.0.0.0 --port 4242Other devices on your network can now reach http://<this-machine-ip>:4242. Linking still requires your master token, so it isn’t open access — but treat it like any service exposed on your LAN.
Reaching it from your phone or a cloud AI
Section titled “Reaching it from your phone or a cloud AI”Devices outside your network — and cloud agents like Claude mobile or ChatGPT — need a stable, reachable HTTPS address. How you provide that is your call: a tunnel (Tailscale, Cloudflare Tunnel, and similar) is the common approach and avoids opening ports on your router. Treeline only needs to be reachable at a URL; the networking itself is standard and up to you.
Good to know
Section titled “Good to know”- The Hub serves the data on this machine — it’s both your everyday Treeline and the point your other devices push to and pull from, working off the same database.
- It’s only reachable while the machine is awake and running. If you want a Hub that’s always available, run an always-on Hub instead.