Ross ROSS = Recommend OSS · open-source software intelligence for agents

yvgude/lean-ctx

Control what your AI can see. LeanCTX (Lean Context) is the context intelligence layer for AI agents — one local Rust binary that decides what they read, remembers what they learn, guards what they touch, and proves what they save. 60–90% fewer tokens as the receipt. 76 MCP tools, 30+ agents, local-first. observed · 2026-08-28

github.com/yvgude/lean-ctx · homepage · Rust · Apache-2.0 (permissive) observed · 2026-08-28

Health v2 · maintenance only

77/100

  • Activity 99
  • Release rhythm 87
  • Longevity 11

Flags: young

How is this computed?

round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10) — computed 2026-09-03. Adoption (stars, forks) is never an input.

  • gap_med: 0
  • age_days: 163
  • days_rel: 7
  • days_push: 7
  • n_releases_24m: 246

Full methodology

Adoption not part of the score

3656 stars · 332 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-29, confidence not recorded

LeanCTX is a local-first Rust binary that acts as a context engineering layer for AI coding agents, providing an MCP server, shell output compression hooks, and persistent project memory. It compresses file reads and command output via tree-sitter AST analysis and caching, claiming 60-90% token savings, and works with 30+ agents like Cursor, Claude Code, and Copilot.

Use cases

  • reduce token costs when using AI coding agents
  • compress file reads and shell output sent to LLMs
  • persist agent memory and decisions across chat sessions
  • coordinate multiple AI agents with shared context
  • monitor and benchmark AI agent token usage
  • set up an MCP server for coding assistants
  • avoid agents re-reading the same files every session

When to choose

  • you use AI coding agents daily and want to cut LLM token spend
  • your agent re-reads files and re-runs commands repeatedly each session
  • you want local-first, offline context tooling where code never leaves your machine
  • you need cross-session memory and multi-agent coordination for coding workflows

When to avoid

  • you need a fully lossless view of all tool output by default (compression is lossy in aggressive modes)
  • you rely on agents or tooling outside its supported 30+ agent integrations
  • you want a hosted or cloud-managed context solution without self-hosting anything
  • your workflow is simple enough that token costs are not a concern

Facets

cli-tool · maturity active

mcp caching compression developer-tools cli monitoring agent-framework developer-tools large-language-models windows cli rust cross-platform context-engineering token-optimization mcp-server ai-coding-agents prompt-compression local-first session-memory tree-sitter cost-reduction ai-agents command-line automation macos linux

10 sources

Member repositories

RepositoryRoleHealth v2
yvgude/lean-ctxmain77

For agents

markdown · JSON · MCP: product_card(name="yvgude/lean-ctx")

Data as of 2026-08-30T08:39:29.467469+00:00 · Report a problem