Open at the core.
Kin, the semantic substrate under KinLab, is open source under Apache-2.0: semantic version control, the transparent VFS, the MCP server, the editor, and the engine. The expensive failure is rarely in the file that changed. Files compile locally; systems break relationally. You can run all of it locally today, with no account and no time limit.
What’s open, what’s hosted
The Kin substrate is Apache-2.0 and runs entirely on your machine. KinLab is the proprietary hosted control plane: everything that only makes sense across many users, many repos, and an org’s shared memory.
Open (Kin)
Apache-2.0Self-hosted, free forever, no account required.
- Semantic version control: the graph-native repo format, CLI, and daemon
- Transparent VFS: graph-backed files served to any tool (Linux + macOS)
- MCP server: semantic tools for Claude, Cursor, Codex, and Gemini
- Editor extension: entity explorer, semantic search, and dependency trace
- The engine: embedded graph storage, lexical and vector retrieval, local inference
Hosted (KinLab)
ProprietaryThe org graph, governance, and team collaboration layer.
- Org graph and memory across every repo and agent
- Governance: bounded, auditable agent access at team scale
- AI semantic review and approval gates for teams
- Hosted collaboration, admin, and enterprise controls
- Cross-repo blast-radius and the shared control plane (the vision)
The open repos
Every surface of the Kin substrate ships in the open. Each one is its own Apache-2.0 repository you can read, fork, and run.
The semantic system of record.
The SQLite of semantic graphs.
Make any tool graph-native, invisibly.
Kin in your editor, not a fork.
Semantic context for any AI agent, over MCP.
The shared vocabulary of the graph.
Language-server intelligence for the graph.
Content-addressed storage substrate.
Fast lexical search under the graph.
Semantic vector search under the graph.
Pure-Rust embeddings, no external dependency.
Reproducible proof, not benchmark theater.
Our stewardship promise
Features that ship open in Kin never move to a paid tier. The moat is the hosted control plane and your org’s graph. Not clawing back features. What is open stays open.
Why open-core
Open-source Kin substrate, hosted KinLab control plane. Adopt the tooling for free, scale on the platform. It is the playbook that worked for the developer platforms we admire: the open substrate is the adoption wedge, and the hosted layer is the business.
GitLab
Open-core DevOps platform: the open product drove adoption; paid tiers added team and enterprise governance on top.
Vercel · Next.js
An open framework everyone runs locally, with a hosted platform that monetizes scale, collaboration, and operations.
Supabase
Open-source primitives you can self-host, plus a managed cloud for teams that would rather not run the infrastructure.
Self-host on-ramp
Install the CLI and daemon, point them at a repo, and Kin parses your code into a semantic graph locally. Wire it into your AI agents with kin setup, then read the source on GitHub.
Adopt the open tooling. Scale on the platform.
The open Kin substrate is free forever. When your team needs the org graph, governance, and hosted collaboration, KinLab is the control plane on top.
Read the reproducible proof, or browse the open Kin ecosystem.