TermCanvas spreads all your terminals across an infinite spatial canvas — no more tabs, no more split panes. Drag them around, zoom in to focus, zoom out to see the big picture, and annotate with freehand drawings.
It organizes everything in a Project → Worktree → Terminal hierarchy that mirrors how you actually use git. Add a project, and TermCanvas auto-detects its worktrees. Create a new worktree from the terminal, and it appears on the canvas instantly.
Canvas
- Infinite canvas — pan, zoom, and arrange terminals freely
- Three-layer hierarchy — projects contain worktrees, worktrees contain terminals
- Live worktree detection — new worktrees appear automatically
- Double-click a terminal title bar to zoom-to-fit
- Drag-to-reorder terminals within a worktree
- Drawing tools — pen, text, rectangles, arrows for annotations
- Workspace save/load — persist your entire layout to a file
AI coding agents
- First-class support for Claude Code, Codex, Kimi, Gemini, and OpenCode
- Composer — a unified input bar that sends prompts to the focused agent, with image paste support
- Live session status — see at a glance whether an agent is working, waiting, or done
- Session resume — close and reopen an agent terminal without losing context
- Inline diff cards — review an agent's changes without leaving the canvas
General terminals
- Shell, lazygit, and tmux terminals live alongside AI agents on the same canvas
Usage tracking
- Token usage and cost dashboard — total spend, per-project breakdown, per-model breakdown
- 24-hour cost sparkline and cache hit/miss stats
Settings & i18n
- English and Chinese (auto-detected from system locale)
- Adjustable terminal font size (6–24 px)
- Auto-update with in-app changelog
CLI
termcanvas— control the canvas from your terminal: add projects, create terminals, read output, view diffshydra— spawn AI sub-agents in isolated git worktrees, then review and merge their work
Download — grab the latest build from GitHub Releases.
Build from source:
git clone https://github.com/blueberrycongee/termcanvas.git
cd termcanvas
npm install
npm run devInstall CLI tools — after launching the app, go to Settings → General → Command line interface and click Register. This adds termcanvas and hydra to your PATH, and installs the Hydra skill into both Claude Code and Codex skill directories.
Both CLIs are bundled with the app. Register them from Settings to use in any terminal.
termcanvas project add ~/my-repo # add a project to the canvas
termcanvas project list # list projects
termcanvas terminal create --worktree ~/my-repo --type claude
termcanvas terminal status <id> # check terminal status
termcanvas diff ~/my-repo --summary # view worktree diff
Hydra lets you break a big task into smaller pieces and hand each piece to an AI agent — Claude, Codex, Kimi, Gemini, or OpenCode. Every agent gets its own git worktree and its own terminal on the canvas, so you can watch them all work in parallel and step in whenever you need to.
Spawn an agent:
hydra spawn --task "fix the login bug" --type claude --repo .
# → returns JSON: { agentId, terminalId, worktreePath, branch, resultFile }This creates a new worktree + branch, opens a terminal on the canvas, and sends the task to the agent. The agent works in full isolation — it can only touch files inside its own worktree.
Pass --auto-approve to inherit the parent agent's permission level (maps to --dangerously-skip-permissions for Claude, --dangerously-bypass-approvals-and-sandbox for Codex).
For read-only tasks (code review, analysis), point to an existing worktree instead of creating one:
hydra spawn --task "audit auth for vulnerabilities" --type claude --repo . --worktree ./my-worktreeMonitor, review, and merge:
hydra list # see all agents and their status
termcanvas terminal status <id> # check if an agent is done
termcanvas diff <worktree> --summary # review what the agent changed
cat <resultFile> # read the agent's summary
git merge <branch> # adopt the changes
hydra cleanup <agent-id> # remove the worktree and terminalSetup:
hydra init # add Hydra usage instructions to your project's CLAUDE.md and AGENTS.mdThis teaches your main AI agent when and how to spawn sub-agents automatically.
| Shortcut | Action |
|---|---|
⌘ O |
Add project |
⌘ B |
Toggle sidebar |
⌘ T |
New terminal |
⌘ ] |
Next terminal |
⌘ [ |
Previous terminal |
⌘ E |
Unfocus / refocus last terminal |
⌘ 1 |
Terminal size: default |
⌘ 2 |
Terminal size: wide |
⌘ 3 |
Terminal size: tall |
⌘ 4 |
Terminal size: large |
On Windows/Linux, replace
⌘withCtrl.
| Layer | Technology |
|---|---|
| Desktop | Electron |
| Frontend | React, TypeScript |
| Terminal | xterm.js, node-pty |
| State | Zustand |
| Styling | Tailwind CSS, Geist font |
| Drawing | perfect-freehand |
| Build | Vite |
Contributions welcome — fork, branch, and open a PR. Licensed under MIT.
