Get Started

LeanCTX documentation

LeanCTX (short for Lean Context) is the context intelligence layer for AI agents. Start with an install, learn by journey, or jump straight to a topic or tool.

The one-paragraph definition

LeanCTX (short for Lean Context) is the open-source context intelligence layer for AI agents. One local Rust binary decides what agents read (10 read modes, 60–90% fewer tokens, ~13-token cached re-reads), remembers what they learn (persistent sessions, knowledge graph), guards what they touch (PathJail, secret redaction, budgets, injection detection) and proves what they save (Ed25519-signed ledger, reproducible benchmark). Compression is one of five subsystems, and every original stays locally retrievable. Works with 30+ AI coding tools via MCP and shell hooks; embeds in any agent via a versioned /v1 API with Python, TypeScript and Rust SDKs. Local use is free forever, enforced by CI.

Read the full story

Start here

Getting Started→ Install the binary and auto-connect every AI tool. Quick Reference→ The commands and tools you reach for daily.

Learn by journey

All 23 journeys →

The twenty-three code-grounded journeys are how LeanCTX is actually used, grouped into four journey tracks. Pick the one that sounds like you.

Get Started→ From zero to compressed reads in ten minutes — auto-detected for your editor. Setup & Onboarding
Daily Workflow→ Compressed reads, persistent memory and smart routing on every turn. Daily Use · Memory & Knowledge · Code Intelligence · Context Engineering & Observability · Analytics, Insights & Reporting · Beyond Coding: Web & Research · Context Firewall
Scale & Teams→ Shared memory, providers and a team server — one brain across many agents. Proxy & Power Integrations · Multi-Agent Collaboration · Team, Cloud & Self-Hosting · Multi-Repo Workspaces · MCP Tool-Catalog Gateway · Build Your Own Agent: SDKs & /v1 API · Context Personas · Universal Intake: Docs, Data & Email · Extend Without Forking: Plugins & WASM
Operate & Govern→ Security, roles, budgets and performance tuning for real codebases. Lifecycle & Troubleshooting · Customization & Governance · Security & Governance · Performance Tuning · Prove It: Savings, Quality & Benchmarks · Per-item Sensitivity Floor

Browse by topic

Core Concepts→ The mechanics behind the savings — read modes, caching, compression — plus how to adapt and extend the engine. 10 pages
Tools→ The MCP tool surface, grouped by what each tool is for. 7 pages
Workflows→ Multi-step patterns: planning, multi-agent and packaged context. 3 pages
Context Engine→ The persistent brain: memory, control and external data sources. 4 pages
Operations→ Run it safely: configuration, security, budgets and tuning. 5 pages
Reference→ The CLI, the API, the SDK surface and the changelog. 3 pages
76MCP tools
10read modes
18languages
95+shell patterns
3integration modes