Features¶
Everything claudectl does, grouped. Each card is a page.
-
Browse, search, tag, fork, resume, archive and export every session across every project — plus usage analytics and per-project launch control.
-
A semantic graph of your codebase, injected through three token-budgeted surfaces, bounded as the project grows, learning durable lessons from every session.
-
An interactive, self-contained HTML view of your project's real dependency structure, expandable from repo down to single files.
-
Plan on an accurate model, execute on a cheap — or completely free — one, from the launcher.
-
Two or more Claude accounts side by side, one row per project, cross-account context injection and account-accurate memory.
-
Start a fresh session seeded with a previous one's transcript — under any account. For when the context window fills up, or an account hits its limit mid-task.
-
Add, remove and inspect MCP servers, see which ones are actually connected, and document their tools.
-
A category-organized agent library, per-project subagent selection, adaptive suggestions, and the skills manager.
-
19 ready-made Claude Code hook templates — formatting, safety guardrails, audit, context injection, token savers — plus AI-generated ones.
-
Status line, failover & checkpoints
A cheap per-turn status line, a proxy that retries a dead model instead of hanging, and a read-only view of Claude Code's checkpoint store.
-
The whole thing as a native desktop app — full parity with the terminal UI, 29 palettes, 7 skins and 4 themed worlds.
Guides¶
- Token economy — how the per-turn context cost is measured and cut
- Project health & auto-fixes — launcher-side mitigations for common Claude Code problems
- Installing the agent library — bulk-install 154 community subagents
Reference¶
- Files, layout & encoding — where everything is stored
- API reference — the local HTTP API the GUI is built on