Kai

Semantic infrastructure for code change.

Kai understands what code means — functions, dependencies, behavior impact — not just which lines changed. This semantic graph powers precise CI, context-aware IDEs, and verifiable AI coding agents.

kailayer.com

Install

# curl
curl -sSL https://get.kailayer.com | sh

# Homebrew
brew install kailayerhq/kai/kai

Quick Start

kai init
kai capture
kai diff                    # semantic change impact
kai ci plan                 # minimal test selection

For full command reference, see docs/cli-reference.md.

MCP Server

Kai ships an MCP server that gives AI coding assistants access to call graphs, dependency maps, impact analysis, and test coverage.

# Claude Code
claude mcp add kai -- kai mcp serve

# Or without installing kai (npx downloads it automatically)
claude mcp add kai -- npx -y kai-mcp

No setup required — the server lazily initializes the semantic graph on first use.

Also available on the MCP Registry as io.github.kailayerhq/kai.

See docs/mcp.md for Cursor setup, tool reference, and troubleshooting.

What Kai Builds

For every commit, Kai constructs a queryable semantic model:

This graph is immutable, verifiable, and designed for machine reasoning.

Use Cases

Selective CI

Kai determines which tests actually need to run based on behavioral impact, not file diffs. Result: 80% CI time reduction for early users.

AI Code Context

11 MCP tools give AI assistants structured access to your codebase's dependency graph, call graph, impact analysis, and test coverage.

Verified AI Agents

Agent proposes edit → Kai validates impact → agent executes with proof, not generation with hope.

Architecture

Kai is open-core: CLI and semantic engine are Apache 2.0. Kai Cloud adds hosted infrastructure and team features.

See What's Open Source vs. Kai Cloud.

Roadmap

Full roadmap →

License

See LICENSE.

Community & Contributing

• Slack
• Contributing guidelines