Burning away Git complexity
Sage wraps everyday Git pain points behind a single, intuitive CLI (and an optional sync server for shared stacks). It keeps branch stacks tidy, automates the boring bits, and gives you a slick TUI when you feel fancy – no magic, no yak shaving, just clean commits.
- 🌱 Elegantly simple – zero mental overhead, readable Rust, no hidden side‑effects
- ✨ DX first – colourful output, progress bars, AI‑assisted commit messages (opt‑in)
- 🪄 Stack aware – restack, navigate, submit for review in seconds
- 🔌 Plugin hooks – extend every lifecycle stage with your own Rust or shell plugins
- 🖥️ TUI – full‑screen dashboard (enable with
--features tui)
# Install sage with all features
just install
# Or install with specific features
just install-only --stack --ai
# Try commands without installing
just try work my-feature
just try-with stack,tui dashNote:
--features stackturns on advanced stacked‑diff commands,--features aienables AI commit message generation.
| Command | Purpose |
|---|---|
work <branch> |
Smart create/checkout |
save [-m|--ai] [--all] <…> |
Stage → commit |
sync [--continue|--abort] |
Restack + push |
share [--draft|--ready] |
Create/update PR |
dash [--watch] |
Repo dashboard |
clean [--remote] [--dry-run] |
Prune branches/reflog |
undo [id] / history |
Revert + log |
resolve |
Launch mergetool |
stats [--since <date>] |
Repo statistics |
doctor [--fix] |
Env/toolchain check |
config <op> |
Manage config |
completion <shell> |
Shell completions |
Enable feature flags for extra stack, ai, and tui commands.
To configure Sage to use your locally running Ollama API for AI-powered commit messages:
# Set the API endpoint (default Ollama port is 11434)
sage config set ai.api_url http://localhost:11434
# Set your preferred model (e.g., llama2, codellama, mistral)
sage config set ai.model gemma3n:latest
# Verify your configuration
sage config get aiNow you can use AI features with your local Ollama instance:
# Generate commit message using local Ollama
sage save --aisage/
├── bins/
│ ├── sage-cli/ # 🌿 main binary
│ └── sage-server/ # stack‑sharing sync service (optional)
├── crates/
│ ├── sage-core/ # domain logic
│ ├── sage-git/ # git2 wrappers
│ ├── sage-tui/ # TUI widgets (optional)
│ ├── sage-plugin/ # plugin SDK
│ └── sage-utils/ # misc helpers
├── docs/
│ └── image.png # project logo
# Quick check if code compiles
just check
# Run tests
just test
# Watch for changes and auto-rebuild
just watch
# Run full CI pipeline locally
just ci
# Build release version
just releaseFor more development commands, run just help.
We believe in using our own medicine. Throughout development Sage manages its own repository—every branch, save, sync, and PR is executed with the CLI you see taking shape here. Expect real‑world polish to land fast because we feel the pain first.
- Fork & clone
just watch– auto-rebuild on changesjust ci– ensures code passes all checks- Submit a PR 🚀
Sage is early‑stage and evolving rapidly. The commands, feature flags, and server behaviour described above are aspirational; not everything is implemented (yet) and details may change without notice. Use at your own risk, and expect breaking changes while we burn away the rough edges.
Licensed under either of
- MIT – see LICENSE-MIT or http://opensource.org/licenses/MIT
- Apache 2.0 – see LICENSE-APACHE or http://www.apache.org/licenses/LICENSE-2.0
at your option.