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.
-
Claude Code plugin: a self-improvement loop that logs user-correction moments, finds patterns, and prescribes recurrence-prevention fixes (MCP server + 2 skills)
MIT -
pdf-editor-w-self-improving-agentgithub.com/crissymoon/pdf-editor-w-self-improving-agentupdated this year
A PDF Editor w/Crypto Sign - 'am adding a self improving AI Agent, Simple Fill & Sign, MCP server, Browsing & Email Automations, File Worker...
MIT -
Recursive Self-Improvement through Evolving Worlds (Dream-RSI) for LLM agents
npm MIT ★ 14 -
agent-reflexion-mcpgithub.com/mdfifty50-boop/agent-reflexion-mcpupdated this year
Agent self-improvement and reflection MCP server — decision logging, pattern analysis, performance tracking
-
MCP Server for self-improving AI coding skills. Observe → inspect → amend → evaluate loop for SKILL.md files. Works with Claude Code, Codex, and Copilot.
MIT ★ 1 -
mcp-agent-vitaminsgithub.com/menkesu/mcp-agent-vitaminsupdated this year
MCP server for Agent Vitamins — daily self-improvement briefs for AI agents. Curated from X, YouTube, LinkedIn, blogs, and GitHub. Free preview + paid full brief with actionable insights.
npm MIT -
Let your coding agent dream: strategies must survive counterfactual replay before they earn trust. An offline self-improvement loop over an evolving world model — no fine-tuning, the LLM stays frozen. Speaks MCP. Ships its Python server inside the package
npm MIT -
Self-improving, self-hosted memory across every AI coding tool, project, and team (Claude Code, Cursor, Copilot, any MCP client). Autoskill proposes new skills from your sessions, so each project improves automatically. Inspectable, grounded, and yours. Built for teams and enterprise. Open source, MIT.
MIT ★ 4 -
Agent-first control plane for PostHog — source-backed analytics, production investigation, and bounded self-healing and self-improving software workflows.
Apache-2.0 ★ 1 -
cogmemgithub.com/writerslogic/cogmemupdated this year
A self-improving, verifiable memory layer for AI coding agents.
pypi Apache-2.0 -
The self-improvement layer for AI agents. Ratel reads your agent's traces, finds what fails, recommends fixes and runs evals on real sessions before you ship. Self-hosted: one app, one Postgres. Includes the TypeScript and Python SDK
npm MIT ★ 563 -
lorekeeper-mcpgithub.com/jessinra/lorekeeperupdated this year
Self-improving MCP memory server for AI agents. One command, no cloud, no config.
pypi Apache-2.0 -
🧠 Ultimate MCP Server - 20 AI tools with autonomous cognition, self-improvement, and multi-modal processing for AI agents and workflows
★ 3 -
Lorekeepergithub.com/Jessinra/Lorekeeperupdated this year
Self-improving memory for AI agents
Apache-2.0 ★ 5 -
forage-mcpgithub.com/isaac-levine/forageupdated this year
Self-improving tool discovery for AI agents. Agents find, install, and learn to use new tools automatically.
npm MIT -
foragegithub.com/isaac-levine/forageupdated this year
Self-improving tool discovery for AI agents. Agents find, install, and learn to use new MCP tools automatically.
MIT ★ 19 -
MCP server for multi-terminal Claude Code orchestration with DAG task management, parallel execution, and self-improvement
npm MIT -
Hybrid neurosymbolic AI using blackboard architecture. Domain-agnostic anomaly detection + cross-domain correlation + local LLM advice + session-based self-improvement. Python 3.12.
MIT -
Verified, reversible, self-improving Vertical Slice Architecture companion for AI coding agents: an MCP server + CLI for 11 stacks
npm MIT -
ThumbGate Pre-Action Checks self-improve from ranked lessons and repeated failures, hard-block detected secret leaks, and block matches in strict mode.
npm MIT ★ 31 -
ThumbGate Pre-Action Checks self-improve from ranked lessons and repeated failures, hard-block detected secret leaks, and block matches in strict mode.
npm MIT -
Self-improving skill memory for coding agents: learn, recall, reinforce, decay — with provenance on every memory and trust only you grant. Local SQLite, no API key, no cloud.
Apache-2.0 ★ 11 -
The full-state-verification layer for AI coding agents: one native MCP server (Rust host, C engine) that turns a repository into measured, grounded, guarded, predictive, self-improving intelligence with provenance for every claim.
NOASSERTION ★ 5 -
Caura (formerly MemClaw) — governed shared memory for AI agent fleets. Multi-agent, multi-tenant, MCP-native. Trust tiers, keystone policies, audit trails, knowledge graph, self-improving retrieval. Apache 2.0.
Apache-2.0 ★ 559 -
browse-mcpgithub.com/that1drifter/browse-mcpupdated this year
Headless-browser MCP server for any MCP-capable agent (Claude, Codex, Gemini CLI, Cursor, ...). Playwright-based, with accessibility-tree refs, Readability, search, annotated screenshots, and a self-improvement feedback loop. Windows, Linux, and macOS.
npm MIT -
Create your best prompts once, improve them with tests, turn them into assistants, and use them from Claude, ChatGPT, Cursor and Open WebUI, on a free local model or your own key. Self-hosted, with a Windows 11 console.
MIT -
Empathetic AI engineering partner with cognitive memory, multi-agent orchestration (10 agents), 132-signal skill routing, and self-improving reflection loops. Built on Model Context Protocol. By Aldo Karendra.
MIT ★ 3 -
91 tools for AI agents. Turn Chrome into an MCP server — inspect DOM, automate clicks, audit security, mock APIs, extract data, download videos, schedule agents, monitor pages, self-improving recipes. Works with Claude Code, Cursor, Windsurf.
npm MIT -
Give any agent self-improving memory you can trust in production. Every correction is kept, shown to your agent when it matters and enforced on Claude Code and Codex. Built-in memory security: notes cannot instruct the agent, and any edit to the record sh
npm MIT -
181 tools for AI agents. Turn Chrome into an MCP server — inspect DOM, automate clicks, audit security, mock APIs, extract data, download videos, schedule agents, monitor pages, self-improving recipes, full API collection management, business intelligence
npm MIT -
Turn web artifacts into tools your team can use. Publish from files or an AI agent, control access, and keep improving the same URL. Self-hosted and open source, with version history, backend primitives and MCP.
MIT ★ 8 -
decision-pathfindergithub.com/haruhunab1320/decision-pathfinderupdated this year
A self-improving decision tree engine for LLM agents — build, execute, and learn from decision paths. Supports Claude, OpenAI, and Gemini with automatic recommendation caching so agents get faster and cheaper over time. Includes an MCP server for Claude C
npm ISC -
Time Estimation MCP Server — 26 tools for PERT, COCOMO II, Monte Carlo, sprint forecasting, context-based estimation, token-to-time mapping, cost estimation, and schedule risk. Self-improving with calibrated intervals and an auto-closing feedback loop. Bu
npm Apache-2.0 -
Open-source self-improving QA agent for software teams. A test harness with memory. Write tests in natural language for web and mobile. agent-qa learns from every run, adapts to UI changes, and catches regressions before you ship.
npm NOASSERTION ★ 857 -
A Research Cortex. A research companion for Claude. Your library, notes & memory stay on your machine; answers cited from your own library; 30+ specialist skills; auto cross-pollination across papers, meetings, ideas, notes & journal; weekly self-review that drafts improvements for your approval. Reasoning runs on the Claude API.
AGPL-3.0 ★ 5 -
A Research Cortex. A research companion for Claude. Your library, notes & memory stay on your machine; answers cited from your own library; 30+ specialist skills; auto cross-pollination across papers, meetings, ideas, notes & journal; weekly self-review that drafts improvements for your approval. Reasoning runs on the Claude API.
AGPL-3.0 ★ 1 -
Vetono update data
93 agentic MCP tools + 49 specialist agents for every major AI CLI — self-learning, no API keys. Veto gives Claude Code, Codex CLI, Gemini CLI, Cursor, Windsurf, Zed, and JetBrains a council of specialist agents, persistent cross-platform memory, a self-learning router, CI/CD gates, and dependency-hallucination + decision-drift guards — all powered by your existing AI subscription. No separate API keys, no extra billing. **Highlights** - **Council debate** — multi-agent GREEN/YELLOW/RED verdicts on risky changes (LLM-backed via MCP Sampling, with deterministic fallback) - **Cross-platform memory & session handoff** — save on Claude, resume on Gemini or Codex; nothing needs re-explaining - **Self-learning router** — improves task routing from recorded outcomes - **Guards & gates** — CI gate, diff review, secrets scan, dependency-hallucination verification, decision-drift enforcement - **Compact mode** — advertise 9 tools instead of 93 and save ~16K context tokens; every tool stays callable **Links**: [GitHub](https://github.com/jigyasudham/veto) · [npm](https://www.npmjs.com/package/@jigyasudham/veto)
smithery -
sendimentgithub.com/leonwangg1/sendimentupdated this year
MCP server giving AI agents persistent, self-cleaning memory across sessions. Forgets the stale, merges duplicates, resolves contradictions.
MIT ★ 2 -
Self-learning loop for Claude Code — corrections and failures become skills, behind a human gate. MCP + hooks, Windows-native, zero background LLM calls.
MIT ★ 7 -
solo-founder-osgithub.com/alex-jb/solo-founder-osupdated this year
Self-evolving 10-agent stack for one-person companies. Cron + eval + reflexion + skill library + PR-gated evolver. pip install solo-founder-os · 513 tests · MIT.
MIT ★ 12 -
Persistent agent identity for AI agents — a hand-edited identity layer, nightly-distilled experience, and lessons injected every session, shared across your whole fleet. Works with Hermes, OpenClaw, Claude Code, Pi, and opencode.
npm PolyForm-Noncommercial-1.0.0 -
Distills your Claude Code / OpenCode sessions — or any text corpus you ingest — into a dated history of how its subject changed, queryable by your agent over MCP.
npm MIT -
The agentic meta-harness — freeze the model, evolve the harness. An open runtime that routes each query to the cost-optimal model, evolves its own harness (planner/context/reviewer/retry/tool/memory/score policy) and autonomously repairs code, then orches
npm MIT