The file-first model is ending. Kin is the first layer of what comes next. Learn more →
Back to Labs

Kin

v0.1.0-alpha.1 Alpha

Semantic Version Control

Git stores text history. Kin understands code.

Kin replaces file-based version control with a semantic graph of code entities and their relationships. Instead of tracking line diffs, Kin tracks functions, classes, types, and how they connect — then serves precise, token-budgeted context packs to AI agents via MCP.

Capabilities

Semantic Graph

Code stored as entities and relationships, not file diffs. Functions, types, and dependencies form a navigable graph powered by KinDB.

Precise Context

Token-budgeted context packs for AI assistants via MCP. Your agent gets exactly what it needs, nothing more.

Identity Tracking

Semantic fingerprints survive renames, moves, and refactoring. Track a function across its entire lifecycle.

Git Interop

Import from Git, export to Git, but Git is not required. Adopt incrementally without disrupting your workflow.

Multi-Language

TypeScript, Python, Go, Java, and Rust as Tier 1 languages. One semantic model across your polyglot codebase.

Semantic Review

Review changed entities and their impact graph, not line diffs. See what a change actually affects downstream.

Quick Start

terminal

# Clone and build from source

$ git clone https://github.com/firelock-ai/kin.git

$ cd kin && cargo build --release

# Initialize in any project

$ cd your-project

$ kin init

  Initialized semantic repository in .kin/

$ kin status

  12 entities tracked, 47 relationships mapped

$ kin trace MyFunction --compact

  MyFunction → calls 3 functions, used by 2 callers

$ kin context MyFunction --budget 4000

  Built context pack: 3,847 tokens (4 entities, 6 relationships)

Architecture

Kin organizes code understanding into four planes. 18 Rust crates in a Cargo workspace, built to be embedded, extended, or run standalone.

Semantic Plane

The source of truth. Entities, relationships, contracts, and semantic changes stored in KinDB, a purpose-built embedded graph engine.

Projection Plane

Source files, Git commits, and docs rendered from semantic state. Files are outputs, not inputs.

Execution Plane

Workspaces, validation runs, and evidence capture. Proves that changes do what they claim.

Control Plane

Reviews, governance, assistant adapters, and benchmarks. Manages quality and trust.

Releases

v0.1.0-alpha.1 March 19, 2026 Alpha Initial Open Source Release
  • + Semantic graph engine (KinDB)
  • + Tree-sitter parsing (TS, Python, Go, Java, Rust, JS, C)
  • + Content-addressable blob store
  • + CLI: init, commit, status, trace, context, diff, review
  • + MCP server for AI context delivery
  • + Git import/export interop
  • + Session workspaces with reconciliation
  • + Benchmark harness for context quality
  • + Semantic fingerprinting for identity tracking
  • + 18-crate Rust workspace architecture

Roadmap

Where Kin is headed

Phase 1 — Foundation

Complete Pre-release

Core semantic engine and data model

  • Semantic graph extraction via Tree-sitter
  • Entity tracking (functions, types, modules)
  • Relationship mapping (calls, imports, inherits)
  • Content-addressable blob store
  • KinDB embedded graph engine

Phase 2 — Intelligence

Complete Pre-release

AI-native context delivery and deep analysis

  • MCP server for AI agent context delivery
  • Token-budgeted context packs
  • Cross-file dependency analysis
  • Change impact prediction
  • Git import/export interop

Phase 3 — Public Alpha

v0.1.0-alpha.1 Current

First open source release — full CLI, MCP, and 18-crate architecture

  • CLI: init, commit, status, trace, context, diff, review
  • Semantic fingerprinting for identity tracking
  • Session workspaces with reconciliation
  • Benchmark harness for context quality
  • 18-crate Rust workspace architecture

Phase 4 — Collaboration

v0.2.0+ Next

Team-scale semantic workflows

  • Semantic merge (conflict resolution at entity level)
  • Semantic branching
  • Multi-repo graph federation
  • Plugin system for custom analyzers
  • Performance optimization for large codebases

Watch the GitHub repository to get notified of new releases.

Watch on GitHub