╭──────────────────────────────────────────╮ ╭──────────────────────────────────────────╮ │ >>>>>>>>>>>>>>> ctx_read >>>>>>>>>>>> │ │ >>>>>>>>>>>>>>> ctx_read >>>>>>>>>>>> │ ┌─────────────┐ │ >>>>>>>>>>>>>>> compress >>>>>>>>>>>> │ ┌──────────┐ ┌─────────────┐ │ >>>>>>>>>>>>>>> compress >>>>>>>>>>>> │ ┌──────────┐ │ 100K tokens │ =========>│ >>> AST >>> cache >>> signal >>> │ =========> │ ~5K tok │ │ 100K tokens │ =========>│ >>> AST >>> cache >>> signal >>> │ =========> │ ~5K tok │ └─────────────┘ │ >>>>>>>>>>>>>>> filter >>>>>>>>>>>> │ └──────────┘ └─────────────┘ │ >>>>>>>>>>>>>>> filter >>>>>>>>>>>> │ └──────────┘ │ >>>>>>>>>>>>>>> dedupe >>>>>>>>>>>> │ │ >>>>>>>>>>>>>>> dedupe >>>>>>>>>>>> │ ╔══════════════╗ ╰──────────────────────────────────────────╯ ╔══════════╗ ╔══════════════╗ ╰──────────────────────────────────────────╯ ╔══════════╗ ║ raw input ║ ─ ─ ─ ─ ─ ─ ─ ─ lean-ctx ─ ─ ─ ─ ─ ─ ─ ─ ─ ─ ║ output ║ ║ raw input ║ ─ ─ ─ ─ ─ ─ ─ ─ lean-ctx ─ ─ ─ ─ ─ ─ ─ ─ ─ ─ ║ output ║ ╚══════════════╝ ╚══════════╝ ╚══════════════╝ ╚══════════╝ ┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄ ┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄ ┌──┐ ┌──┐ ┌──┐ ┌──┐ ┌──┐ ┌──┐ ┌──┐ ┌──┐ ┌──┐ ┌──┐ ┌──┐ ┌──┐ ┌──┐ ┌──┐ ┌──┐ ┌──┐ ┌──┐ ┌──┐ ┌──┐ ┌──┐ ┌──┐ ┌──┐ ┌──┐ ┌──┐ ┌──┐ ┌──┐ │░░│ │░░│ │░░│ │░░│ │▓▓│ │▓▓│ │▓▓│ │░░│ │░░│ │░░│ │░░│ │░░│ │░░│ │░░│ │░░│ │░░│ │░░│ │▓▓│ │▓▓│ │▓▓│ │░░│ │░░│ │░░│ │░░│ │░░│ │░░│ │░░│ │░░│ │░░│ │░░│ ───> │▓▓│ │▓▓│ │▓▓│ ───> │░░│ │░░│ │░░│ │░░│ │░░│ │░░│ │░░│ │░░│ │░░│ │░░│ ───> │▓▓│ │▓▓│ │▓▓│ ───> │░░│ │░░│ │░░│ │░░│ │░░│ │░░│ └──┘ └──┘ └──┘ └──┘ └──┘ └──┘ └──┘ └──┘ └──┘ └──┘ └──┘ └──┘ └──┘ └──┘ └──┘ └──┘ └──┘ └──┘ └──┘ └──┘ └──┘ └──┘ └──┘ └──┘ └──┘ └──┘ file file file file read cache send tok tok tok tok tok tok file file file file read cache send tok tok tok tok tok tok ╭──────────────────────────────────────────╮ ╭──────────────────────────────────────────╮ │ >>>>>>>>>>>>>>> ctx_read >>>>>>>>>>>> │ │ >>>>>>>>>>>>>>> ctx_read >>>>>>>>>>>> │ ┌─────────────┐ │ >>>>>>>>>>>>>>> compress >>>>>>>>>>>> │ ┌──────────┐ ┌─────────────┐ │ >>>>>>>>>>>>>>> compress >>>>>>>>>>>> │ ┌──────────┐ │ 100K tokens │ =========>│ >>> AST >>> cache >>> signal >>> │ =========> │ ~5K tok │ │ 100K tokens │ =========>│ >>> AST >>> cache >>> signal >>> │ =========> │ ~5K tok │ └─────────────┘ │ >>>>>>>>>>>>>>> filter >>>>>>>>>>>> │ └──────────┘ └─────────────┘ │ >>>>>>>>>>>>>>> filter >>>>>>>>>>>> │ └──────────┘ │ >>>>>>>>>>>>>>> dedupe >>>>>>>>>>>> │ │ >>>>>>>>>>>>>>> dedupe >>>>>>>>>>>> │ ╔══════════════╗ ╰──────────────────────────────────────────╯ ╔══════════╗ ╔══════════════╗ ╰──────────────────────────────────────────╯ ╔══════════╗ ║ raw input ║ ─ ─ ─ ─ ─ ─ ─ ─ lean-ctx ─ ─ ─ ─ ─ ─ ─ ─ ─ ─ ║ output ║ ║ raw input ║ ─ ─ ─ ─ ─ ─ ─ ─ lean-ctx ─ ─ ─ ─ ─ ─ ─ ─ ─ ─ ║ output ║ ╚══════════════╝ ╚══════════╝ ╚══════════════╝ ╚══════════╝ ┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄ ┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄┄ ┌──┐ ┌──┐ ┌──┐ ┌──┐ ┌──┐ ┌──┐ ┌──┐ ┌──┐ ┌──┐ ┌──┐ ┌──┐ ┌──┐ ┌──┐ ┌──┐ ┌──┐ ┌──┐ ┌──┐ ┌──┐ ┌──┐ ┌──┐ ┌──┐ ┌──┐ ┌──┐ ┌──┐ ┌──┐ ┌──┐ │░░│ │░░│ │░░│ │░░│ │▓▓│ │▓▓│ │▓▓│ │░░│ │░░│ │░░│ │░░│ │░░│ │░░│ │░░│ │░░│ │░░│ │░░│ │▓▓│ │▓▓│ │▓▓│ │░░│ │░░│ │░░│ │░░│ │░░│ │░░│ │░░│ │░░│ │░░│ │░░│ ───> │▓▓│ │▓▓│ │▓▓│ ───> │░░│ │░░│ │░░│ │░░│ │░░│ │░░│ │░░│ │░░│ │░░│ │░░│ ───> │▓▓│ │▓▓│ │▓▓│ ───> │░░│ │░░│ │░░│ │░░│ │░░│ │░░│ └──┘ └──┘ └──┘ └──┘ └──┘ └──┘ └──┘ └──┘ └──┘ └──┘ └──┘ └──┘ └──┘ └──┘ └──┘ └──┘ └──┘ └──┘ └──┘ └──┘ └──┘ └──┘ └──┘ └──┘ └──┘ └──┘ file file file file read cache send tok tok tok tok tok tok file file file file read cache send tok tok tok tok tok tok

Lean Cortex

The Cognitive Context Layer for Agentic Systems.

LeanCTX helps AI agents perceive, compress, remember, and route context. Less noise, better signal. One binary, zero config, open source.

Understand Your Context See What You Save Star on GitHub

- installs

- GitHub stars

61 AI tools supported

[x] [-] [o] lean-ctx session

$ lean-ctx read src/lib/auth.ts -m map

→ exports: authenticate(), validateToken(), refreshSession()

→ 4,200 → 180 tokens (96% saved)

→ cached: 13 tokens on re-read

$ lean-ctx grep "authenticate" src/

→ 3 matches in 0.8ms

→ auth.ts:14, middleware.ts:8, routes.ts:23

$ lean-ctx -c "cargo build --release"

→ Compiling lean-ctx v3.6.8 in 3.2s

→ compressed: 847 → 42 tokens

curl -fsSL https://leanctx.com/install.sh | sh

The Problem

You can't optimize what you can't see.

Every file read, every shell command, every search result — your AI sends it all to the model without filtering. You have no visibility into what's eating your context window. The result? Wasted tokens, lost focus, and decisions based on noise.

"Compression saves tokens. But knowing what's in your context — that's what changes the output."

Without lean-ctx

AI reads 800-line file

→ Thousands of tokens sent to LLM

Includes comments, whitespace, imports

→ Slow, expensive, hits context limits

With lean-ctx

AI reads 800-line file

→ lean-ctx compresses via AST parsing

Keeps types, signatures, logic

→ 60–90% fewer tokens, same signal

How it Works

Understand, compress, remember, verify, connect.

LeanCTX sits between your AI and your codebase. First it shows you what's in your context — then it optimizes the full lifecycle from file read to model delivery.

Understand

See every file, token count, rule cost, and conversation weight in real-time. Know exactly what's eating your context window before making decisions.

Compress

10 read modes with Tree-sitter AST parsing for 21 languages. Strips noise, keeps structure. Cached re-reads cost ~13 tokens — 60-99% fewer tokens per read.

Remember

Cross-session memory, knowledge graphs, and context checkpoints. Your AI remembers decisions, discoveries, and file relationships across sessions.

Verify

Profiles, roles, and budgets. Control what each agent sees, how much it spends, and when to throttle. Full audit trail.

Connect

Drop a TOML file, connect any API. GitHub, Jira, databases, or your own services — lean-ctx pulls live context from any data source into your AI's working memory.

OWASP Agentic Top 10 aligned

OS-level sandboxing

Full audit trail

See What You Save Learn about shell patterns

See It In Action

What it looks like in your terminal.

lean-ctx works silently behind the scenes. When your AI reads a file, the context layer transparently compresses, routes, and optimizes the output. Here's a real ctx_read call and the metrics dashboard.

[x] [-] [o] ctx_read - map

AI calls: ctx_read({ path: "src/lib/mcpManifest.ts", mode: "map" })

schema_version 1

tools.granular 61

tools.unified 5

read_modes 10

modes: auto, full, map, signatures, diff, aggressive

Real-time context visibility — see every file, token count, and rule cost as it happens

60–99% typical token reduction per file read. Real-time visibility into every token. Up to 99% on cached re-reads. See methodology

61 context intelligence tools — compression, visibility, memory, control, and multi-agent support

0 configuration required — install and go. Auto-adapts to your IDE.

Compatibility

Works with every major AI coding tool.

Cursor, Claude Code, GitHub Copilot, Windsurf, Pi, Neovim, Sublime Text, Emacs, Codex — lean-ctx integrates with all of them. Zero vendor lock-in, zero configuration per tool. Install once, benefit everywhere.

AiderAmazon QAmpAntigravityAWS KiroClaude CodeClineContinueCursorEmacsGemini CLIGitHub CopilotJetBrainsNeovimOpenAI CodexOpenCodePiQwen CodeRoo CodeSublime TextTraeVerdentWindsurfZed

Full compatibility matrix with setup guides

FAQ

Frequently Asked Questions

What is LeanCTX?

LeanCTX is the Context Layer for AI development. It sits between your AI coding tool and your codebase, compressing file reads by up to 99%, persisting memory across sessions, and verifying all outputs before delivery. It works with 29+ AI tools including Cursor, Claude Code, and GitHub Copilot.

How much does LeanCTX save in token costs?

Active developers save $30-100+ per month on AI API costs. File reads are compressed by 60-99%, shell output by 60-90%, and cached re-reads cost only 13 tokens. Use lean-ctx gain to measure your personal savings.

Does LeanCTX work with Cursor / Claude Code / Copilot?

Yes. LeanCTX supports 29+ AI coding tools out of the box. Run lean-ctx setup and it auto-detects and configures all installed editors. It supports two integration modes: Hybrid (MCP + shell hooks) and Full MCP.

Is LeanCTX free and open source?

Yes. LeanCTX is fully open source under MIT + Apache-2.0 dual license. Zero telemetry, everything runs locally. The single Rust binary is available via npm, cargo, or direct download.

What is the Context Layer?

The Context Layer is lean-ctx's architecture for managing the full AI context lifecycle. It sits between your code and the AI: deciding what to read, how to compress it, what to remember, and how to verify results. Learn more on the Architecture page.

Stop guessing. Start seeing what's in your context.

One Rust binary. No cloud. No account. MIT + Apache-2.0 licensed. Install in 60 seconds — see what's in your context, then optimize it.

Get Started Star on GitHub