sail ⛵️

sail is an IDE built for agents, not humans.

AI coding usually breaks down for the same reason human teams do: the codebase gets noisy, inconsistent, and hard to navigate. sail fixes that by exposing an agent-first interface that gives models the right context at the right time and enforces good standards on every write. The result is a more predictable, more scalable codebase that agents can keep working in without drifting into chaos.

Just add sail into your project and prompt your agent to code through sail and see the magic instantly:

Sail guides your agent towards writing quality code by default:

Sail blocks your agent when it breaks conventions. Tech debt must be paid immediately!

Quick Start 🚀

Web Apps 🌐

React apps, Next.js apps, and other browser-based projects.

npm install -g sail
npx create-react-app my-app
cd my-app
sail init
claude "build me a chess game with AI player support"

Standalone Apps 🛠️

Scripts, CLIs, and other non-web TypeScript projects.

npm install -g sail
sail init my-app
cd my-app
claude "build me a reddit CLI"

What It Is ✨

sail gives agents a development environment designed around structure, constraint, and fast feedback.

Instead of exposing a free-for-all filesystem, sail gives the agent a tighter interface for exploring code, making changes, and getting immediate feedback when it starts to drift. That makes it easier to keep an AI-driven project coherent as it grows.

Some Awesome Things Sail Does

• 🧭 Gives your agent an easy way to trace all dependencies and usages of any piece of code it writes, so it can manage context without getting lost.
• ✅ Does a compile and type check on every agent edit, giving your agent immediate feedback on broken imports, bad types, and invalid code. No more silently broken projects.
• 🧪 Forces your agents to write an appropriate number of tests for every file right after making writes or edits, instead of leaving the test suite as "future work."
• 🧱 Keeps your public boundaries clean, so agents do not slowly turn your app into a bowl of spaghetti.

How Sail Works ⚙️

sail organizes a TypeScript project as a graph of named nodes. Agents explore that graph through commands like query, read, and graph, then make changes through write, patch, and test commands instead of editing files directly. Every mutation is checked by the tool, so structure, validation, and test expectations are enforced as the project evolves.

Why Use Sail 🧠

• Agents get the right context at the right time instead of scanning broad file trees.
• Important rules live in tooling, not just in prompts that models can forget.
• All written code is validated immediately, so bad changes fail fast.
• The codebase stays more predictable and structured as it grows.
• You get better results from AI without constantly cleaning up architectural drift.