MCP servers that ship as code — a repository or a package you install and run locally (usually over stdio). They have no hosted endpoint, so we cannot probe or grade them: instead each one shows its repository, install command, license, stars and when it was last updated.
Collected from GitHub repositories and npm, PyPI, Cargo, Go, Maven, NuGet and RubyGems packages tagged as MCP servers, the official MCP Registry, Docker's MCP catalog, ModelScope, Baidu MCP World, Cline and other public directories; one entry per repository. A server that also offers a hosted endpoint is listed under Remote instead. These entries are not counted in our reachability, verification or readiness statistics.
-
A token-lean MCP memory server for AI coding agents: a knowledge graph your agent can search without reading the whole thing. Drop-in replacement for the reference memory server.
npm MIT ★ 3 -
LeanKG: Stop Burning Tokens. Start Coding Lean.
Apache-2.0 ★ 222 -
Fast Lean 4 proof feedback for coding agents. CLI, Python library, and MCP server with warm LeanInteract sessions and cached env reuse.
MIT ★ 7 -
Ultra-lean, scope-aware code search for LLMs and developers
npm Apache-2.0 ★ 1 -
claude-code-sshgithub.com/hunchom/claude-code-gitupdated this year
MCP server giving Claude Code 51 typed SSH tools across your server fleet. Pooled connections, per-user gated, context-lean output.
npm MIT -
Code intelligence for agents: find the code that matters and keep your context window and tokens lean.
Apache-2.0 ★ 163 -
claude-code-sshgithub.com/hunchom/claude-code-sshupdated this year
MCP server giving Claude Code 51 typed SSH tools across your server fleet. Pooled, gated, context-lean — stop being the middleman.
MIT ★ 2 -
Fast, lean MCP server that lets Claude Code and Codex ask your Google NotebookLM notebooks questions: cited answers, without loading your docs into context.
★ 1 -
Drive a live Blender with Claude: in-Blender chat on your subscription + token-lean MCP ops for Claude Code
GPL-3.0 ★ 4 -
Turn PDFs, Office files and scans into token-lean Markdown for LLMs and coding agents. Local-first: web app, desktop, CLI, MCP server, Claude Code & OpenCode hooks.
Apache-2.0 -
Generate, preserve, and verify lean AGENTS.md, RUNBOOK.md, and machine-readable repo context for AI coding agents.
MIT ★ 5 -
vibecoding-with-closed-eyes-on-sofagithub.com/PennyBurden/vibecoding-with-closed-eyes-on-sofaactive
静默改代码，张嘴说人话 · Close your eyes, lean back, and let your AI agent code while you talk like a boss 🛋️
MIT ★ 1 -
Connect Claude Code to AgentsChat — AI Agent social network. Core tools stay lean while extended tool groups load on demand for lower token overhead and cleaner role-specific context.
npm Apache-2.0 -
agentchat-mcpgithub.com/swswordholy-tech/iosdevupdated this year
Connect Claude Code to AgentChat — AI Agent social network. Core tools stay lean while extended tool groups load on demand for lower token overhead and cleaner role-specific context.
npm MIT -
The model proposes, deterministic code disposes, a human signs. Content-addressed envelopes, hash-chained audit, signed approval — for ledgers, Lean proofs, manuscript claims, reproductions, benchmarks, peer review. Conformance-tested; byte-compatible with @kernel.chat/kbot-finance.
MIT -
LeanCTX is the Context Gateway for AI Systems. Select task-relevant context, apply supported controls, connect local knowledge and inspect recorded evidence. Open-source Engine: CLI, MCP, hooks and supported proxy paths.
pypi Apache-2.0 ★ 3,951 -
lean-local-search-mcpgithub.com/dududuguo/lean-local-search-mcpupdated this year
MCP server for fast local Lean and Mathlib theorem search with tree-sitter indexing and proof probes
MIT ★ 1 -
Self-hosted Linear MCP server with response-shape control — lean field-trimmed reads, minimal write acks, drop-in tool names. Built for token-frugal LLM agents.
npm MIT -
A token-lean MCP memory server: a knowledge graph your AI agent can search without reading the whole thing. Drop-in replacement for @modelcontextprotocol/server-memory.
npm MIT -
Index codebases into a local knowledge graph to provide accurate context for AI coding tools through a direct MCP server connection.
★ 1 -
High-performance WebAssembly theorem prover with dependent types, hash-consing (150x faster), Ed25519 proof signatures, MCP support for Claude Code, AgentDB vector search, episodic memory, and ReasoningBank learning. Formal verification with cryptographic
npm Apache-2.0 -
A native desktop IDE for Lean 4 on macOS, Windows and Linux. Every proof re-checked by Tenet, an independent Lean 4 kernel. An MCP server for AI assistants, with Git and GitHub built in.
MIT ★ 8 -
Plan, apply and roll back changes to a production ontology, with a blast radius report and a proof an auditor can re-check.
pypi MIT ★ 2,904 -
telosgithub.com/nocoder26/telosupdated this year
The AI that asks why. A Claude Code plugin that refuses to let founders write product code until they've validated with real customers. 13 skills (concierge onboarding, adversarial customer roleplay, optional gbrain bridge) + brain MCP. Free, MIT.
MIT -
Unofficial lean MCP server for Yandex Tracker: 9 tools, ~1.2k tokens of tool definitions (official: ~24.8k), compact answers, one-command setup.
MIT -
contextive-mcp-servergithub.com/BjornMelin/contextive-mcp-serverupdated this year
A lean, multi-provider MCP server that unifies LLM access. Token-efficient tools, composable workflows, and multi-model routing via Vercel AI SDK. Works with Claude, ChatGPT, Cursor & more.
-
Zero-trust L0-L5 verification gateway in Zig 0.17: native Z3 proof-DAG checker, Belnap B4 logic, 42 Lean 4 theorems (0 sorry), 8-lane SIMD BLAKE3 (.kimg) & MCP server.
NOASSERTION ★ 13 -
🦛 A hippocampus for your AI. Long-term memory for Claude & any MCP client, stored as Markdown in your own GitHub repo: hybrid semantic search, knowledge graph, auto-memory, compaction-proof, token-lean.
MIT ★ 1