Set up Kin

Install the CLI, pick your path, connect a repo, and prove your local setup is healthy.

first-run setup

1. Choose your platform

Install commands and notes adapt to your operating system.

Install with the one-line installer
curl -fsSL https://get.kinlab.dev/install | sh

The Kin core supports Apple Silicon and Intel. VFS projection uses DYLD injection and may be blocked by SIP or hardened executables.

2. Choose your path

Local CLI, editor, agent/MCP, hosted connect, or advanced installs.

Initialize a repository and run semantic queries directly from your terminal.

  1. 1Initialize the semantic graph for the current repository.
    kin init
  2. 2Run a graph + lexical locate immediately; vectors are not required.
    kin locate "where is auth handled"
  3. 3Optional: add vector similarity with the roughly 522 MB default model; budget about 2 GB of available memory for public-alpha use.
    kin embed
  4. 4Build an AI-native context pack for a target entity.
    kin context <entity>

Prove your setup is healthy

Run the command locally and paste the JSON output. KinLab parses the real report. Nothing is marked healthy unless your own kin setup status says so.

kin setup status --json
Expected checks
  • kin binary
  • kin-daemon binary
  • VFS projection
  • Repository
  • Shell integration
  • AI client MCP config
  • Editor extension
  • KinLab connection
  • Semantic query readiness