Zaguán Blade is an AI-native code editor built from scratch with Rust as the engine. It separates the fast native UI from the AI daemon, keeps your code on your machine, and supports local Ollama workflows or cloud models when you need speed.
Built for Control: Blade is designed for engineers who want transparent AI workflows, Git-backed review, predictable file changes, low resource usage, and no opaque background magic.
Modern AI editors can feel bloated, opaque, and cloud-dependent. Zaguán Blade takes a different route: a Rust-first desktop editor that treats Git and files on disk as the source of truth.
AI changes are applied transparently, tracked with snapshots, reviewed in diffs, and accepted or rejected through a workflow you can understand. The goal is not magic; it is controlled, inspectable AI development.
Run locally with Ollama or connect to OpenAI-compatible endpoints and cloud providers when you need additional speed or capability.
AI writes to disk with history snapshots. Accept keeps the change; reject restores from the tracked snapshot.
AI can read, write, edit, search, run approved terminal commands, inspect Git state, and work across multi-file tasks.
Loop detection, stagnation detection, tool-spam prevention, and parallel reads help prevent runaway workflows and wasted tokens.
Review pending AI changes with diffs, accept or reject per file, and batch accept or reject groups of changes.
Multiple persistent terminal instances with command approval workflows and AI-readable output for debugging loops.
Tree-sitter parsing and SQLite full-text search power fast symbol lookup across Rust, TypeScript, JavaScript, Python, TSX, and JSX.
Visual diffs, history, branch workflows, staging, and AI-assisted changes stay aligned with Git as the review boundary.
Tauri v2, Vite, React 19, CodeMirror 6, TailwindCSS 4, xterm.js, and system WebView for a fast native desktop shell.
Tokio, tree-sitter, rusqlite, diffy, notify, reqwest, and WebSocket clients handle performance-critical editor operations.
A dedicated AI daemon manages state, tool execution, model integration, and the long-running agentic workflow loop.
Zaguán Blade v0.6.1 is available from GitHub releases for Windows, macOS, and Linux. Current packages include Windows MSI/EXE, macOS DMG/TAR.GZ for Apple Silicon, and Linux DEB/RPM/AppImage builds.
# Build from source with pnpm
pnpm install
pnpm tauri dev
# Production build
pnpm tauri buildActive Development: Zaguán Blade is evolving rapidly toward v1.0.0. Expect frequent updates, a fast feedback loop, and occasional breaking changes while the editor and daemon mature.
Zaguán AI: The hosted/cloud workflow uses Zaguán AI subscription access, while local workflows can run through Ollama or compatible local endpoints.