TheBrain API: Now Fully Local and Unrestrained
TheBrain now exposes a local HTTP API directly from the desktop app. Featuring blazing speed, real-time UI updates, and uncompromising power.
Since we introduced TheBrain API in the cloud the response has been fantastic, and developers have built things we hadn't imagined. But there was one consistent piece of feedback from everyone pushing the API hard: I want this to be faster, and I want it to work against the brains on my own machine.
Now, as part of the latest version of TheBrain 15, we're delivering exactly that.

What's New
The local API speaks the same shape as the cloud API, so most existing code ports over with little more than a base-URL change. But running in-process next to your brain unlocks a different class of automation:
- Dramatically faster. No network round-trip, no cloud queue. Scripts can create, link, and modify thousands of thoughts in seconds — we've measured workloads sustaining hundreds of thoughts per second end-to-end. Bulk imports that used to take an afternoon now finish before you can switch tabs.
- No rate limits. Because everything runs locally, there's no throttle and no per-minute ceiling. Hammer it as hard as your machine will go.
- Works on any brain you have locally — even closed ones. You're not restricted to the brain currently open in the foreground. Point a request at any brain that exists on the device and the API will operate on it directly. Great for batch jobs that sweep across multiple brains, or for tools that need to read from a reference brain while you work in another.
- Real-time feedback in the running app. When a script creates or modifies a thought, creates a note or adds an attachment, you see it appear immediately in the plex. This instant feedback feels a little like magic.
The one thing to know up front: the desktop app has to be running. That's the trade-off for everything above. If you need always-on availability without a desktop session, the cloud API at api.bra.in is still the right tool. For everything else — local agents, fast batch jobs, AI integrations sitting next to your brain on the same machine — the local API is what you've been waiting for.
A working example: the "Send to TheBrain" browser extension
To show what the local API can do, we've built and open-sourced a Chrome extension called Send to TheBrain. One click from any web page captures the page into the currently open brain.
It's available instantly - no syncing and no BrainBox needed.
The full source is on GitHub and is intended as a starting point. Fork it, gut it, copy the bits you need — it's a working reference for authentication, command invocation, and thought creation against the local API.
If you just want to use Send to TheBrain, you can also get it on the Chrome Web Store.
What you can build
Just a few ideas of the possibilities…
- AI agents that actually live with your brain. Local LLMs and agent frameworks can read context and write back results without ever leaving the machine.
- Capture endpoints from anywhere. Watch folders, email rules, voice triggers, journaling apps, screenshot tools — anything that can issue an HTTP request can land thoughts straight into your brain.
- Bulk migration and shape-changing. Importers, restructurers, taxonomy refactors, and one-off cleanup scripts that would have been impractical at cloud-API speeds.
- Exporting an publishing. Your brain can now serve as the main source for projects such as web sites, lists and anything else you want. This lets you take advantage of TheBrain's infrastructure for sophisticated editing tools, syncing and more, then leverage that data wherever you want.
Getting started
Update to 15.0.534 or later, open the desktop app, and the local API is ready to go. To get the local connection, including documentation, just open Settings > User > Local API.
Happy automating.
More posts by: Harlan Hugh








