tbrain icon

tbrain

One dashboard for every AI coding session on your Mac — Claude Code, Codex, Cursor, OpenCode and Antigravity, all in one place.

Download for macOS

Free · macOS 14 or later · Apple Silicon

What it does

Why it exists

Working across several AI coding tools means the context of what you were doing is scattered across several tools' own session storage, each with its own format and its own place to look. tbrain reads all of it and keeps a single, cross-tool timeline per project, so switching agents doesn't mean losing track of what the last one did.

Using it

Open the app once — it starts its own local hub in the background (a small server and a file watcher, both on your machine only) and remembers it across launches. Pick a project folder, and its sessions, rules, skills and memory show up as soon as it has any.

Click a tool's button at the top to start it against that project in an embedded terminal, right inside the window. Switching tabs doesn't kill the process — it keeps running until you close its tab.

What stays on your machine

tbrain only reads the session folders each tool already keeps on disk — ~/.claude, ~/.codex, ~/.cursor, and the equivalents for OpenCode and Antigravity. Everything it collects goes into one SQLite database under ~/.tbrain on the same machine.

Nothing is sent anywhere unless you turn on the optional AI summarizer and supply your own API key — in that case, the text of a session goes to whichever provider you configured, using your key, and nowhere else. With no key configured, tbrain makes no network calls at all.

What you need

macOS14 (Sonoma) or later, Apple Silicon
Useful withClaude Code, Codex, Cursor, OpenCode, or Antigravity — at least one installed

The app isn't notarized. macOS will warn about an unidentified developer the first time you open it — right-click the app, choose Open, and confirm once.

tbrain is free and comes with no warranty. It's a tool I built to keep track of my own AI coding sessions and published because it might be useful to someone else running the same mix of tools.