Thanks for checking out CCC! Quick context — I built this because every parallel-session orchestrator I tried wanted to OWN execution. The moment you opened a terminal and `claude --resume`d something, the dashboard went blind. CCC inverts it: it reads on-disk state directly, so every AI coding session on your Mac shows up regardless of how it was launched. v3 (today) extends the same bet across engines and repos: → Multi-engine spawning. Claude, Codex (`codex exec`), and Gemini CLI all appear as first-class kanban cards with their own engine chip. Codex is fire-and-watch; Gemini gets full discovery / transcript / spawn / resume parity with Claude. → Multi-repo. A vertical repo sidebar shows every known repo (running CCC servers up top, switchable repos below). The "All repos" view aggregates every conversation across every folder you've ever worked in. → History search. A new 🔎 drawer (or `/` shortcut) does BM25 keyword search over every transcript. Inline sidebar search now also surfaces matches from other repos as you type. → Side-by-side conversations. Drag a session row onto the right or bottom edge of the open conversation to split the pane. Each pane has its own composer and SSE stream. → In-UI terminal. ⌨ button (Cmd/Ctrl+\`) opens a one-shot terminal with cwd clamped to the selected repo. Don't run on untrusted networks. → launchd service install. `./run.sh --install-service` keeps CCC running at login. → Cold scan on big repos: ~135s → ~6.6s. Plus a breaking change: `/api/repo/switch` no longer mutates server state; every repo-scoped endpoint takes explicit `repo_path` / session-derived context now. Full migration notes in the v3 release. Open source, MIT, no telemetry. Happy to answer anything. Methodology write-up: amirfish.ai/parallel-sessions Release notes: github.com/amirfish1/claude-command-center/releases/tag/v3.0.0

Command Center for Claude/Codex/Gemini

Multi-engine Mgr for every AI coding session on your Mac

Multi-engine Mgr for every AI coding session on your Mac