Local-first MCP context engine for AI coding agents — cuts token cost ~90% via a per-project code graph + code folding. Works with Cursor, Claude Code, Codex & any MCP client. Web + ML/PyTorch codebas
Local-first MCP context engine that cuts AI coding-agent token cost by ~90% — a per-project code graph + folding for Cursor, Claude Code, Codex, and every MCP client. Built for web and ML codebases. Status: early — building P0: project isolation on top of the NeuroMesh v0.9.0 baseline.
Derivative of NeuroMesh by yoosef alipour (MIT). See NOTICE · ATTRIBUTION.md · ROADMAP.md.
Upstream NeuroMesh README (baseline v0.9.0)
Local-first MCP context engine for Cursor, VS Code, Claude, Codex, and every MCP client. NeuroMesh indexes your repo into a graph, routes your prompt to the right symbols, and sends a folded evidence packet — not thousand-line file dumps.
Cursor · VS Code · Claude · Codex · OpenCode · MiMo CLI · Antigravity · Kilo · Trae · Windsurf · Zed
The pain · Fold · Galaxy · Measured · Install · Connect · Docs · Site
You ask a simple question in a large project. The editor copies two or three thousand-line files and ships them to the model.
1. Tokens you never needed — dollar cost on every turn 2. Seconds of fake loading while the window fills with helpers you will not touch 3. Lost in the middle — the model drowns in unrelated bodies and invents bugs
Today’s workarounds all leak in a different place:
Approach What goes wrong ----------------------- ------------------------------------------------------------------- Vector RAG Chunks smash functions. The shape of the code disappears. “Just attach the files” The model sees everything and understands nothing. A static code graph Better map — then it still pastes full files into the prompt.
From the project README.
Add the radar badge to your README — it shows your project was picked up by MCP Radar and links to this page:
[](https://mcp.liqiwa.com/s/ParsaVictor--code-context-engine.html)
Keep the signal. Cut the noise. Local CLI + MCP — fold verbose tool output for Claude Code, Codex, Cursor & Desktop.
coffeecoproject/layermapLayerMap MCP: a layered map of your codebase for coding agents: callers, callees and the impact of a change, for TypeScript, JavaScript, Go, Python and Java. MCP server, CLI, and plugins for Claude Co
hung12ct/culiSelf-improving context & memory for Claude Code and OpenAI Codex. One canonical knowledge store injected only when relevant — hooks push budgeted context, an MCP server serves depth on demand, and it
abhishekswe/agent-fastpathJev MCP server: a decision layer for coding agents, built on TypeSafe Jev (System One model). Ship gates, risk checks, file triage that keeps files out of context, and a safe headless browser, with ca
Farukes/TokenJarZero-cost token optimization engine for AI coding assistants (Claude Code, Cursor, Antigravity, Windsurf). Saves 70-95% tokens via AST skeletons, differential cache, and line slicing.
LukeHawkins/turbofigLet AI drive Figma. Instant, lightweight and free: no paid Figma seat, no quota, no Figma token. Open source, for macOS.
The top new MCP servers of the week, every Monday. No spam, unsubscribe anytime.