MCP servers tagged llm
10K live MCP servers tagged "llm", ranked by trust score. Tags come from package metadata and repository topics, so this list covers servers across every category that work with llm.
Intent Compiler
Natural Language → Structured Intent IR → Any Output. The LLVM for vibe coding. 把模糊的自然语言编译成AI Agent能精确执行的结构化意图。
Boule
Lean, research-backed multi-LLM council with bias-aware judge controls (position-swap, verbosity-norm, stake-free judge) adapted from LLM-eval research.
Skillguard
Static security scanner for AI agent skills and plugins — flags prompt injection, hidden Unicode, exfiltration, and dangerous commands in SKILL.md/CLAUDE.md and scripts.
Oghma
Persistent memory for AI coding assistants. Mines transcripts, deduplicates, makes them searchable.
Claude Code Hooks Mastery
🔧 Master Claude Code hooks to control behavior and enhance agent orchestration with sub-agents and validation systems for effective coding solutions.
Tokenpak
Local proxy that puts the right project context in front of the model before it starts working — with a receipt for what was added
Capability Os
AI-powered personal OS with autonomous agents, progressive security, custom agent creator, and multi-channel messaging
Promethea Agent
Local-first cognitive agent runtime with personalized long-term graph memory, reasoning workflows, tools, and Web UI.
Weave Compose
Platform-aware skill composition layer for AI coding tools
Vibe Driven Dev
Agent-first framework for safe vibe coding. Turn messy product ideas into structured plans, bootstrap files, stack decisions, AI model choices, and handoff-ready workflows for cod…
Pastoon
Auto-convert clipboard JSON to TOON format - 40% fewer LLM tokens. macOS tray + MCP server.
Contextual Memory Architecture
A local-first, Obsidian-compatible memory layer for persistent AI agents. Reasoner + Retriever + Recorder over a markdown vault.
Kosh
Kosh is a token elimination and context efficiency platform for AI agents, reducing context retransmission through leasing, batching, and reusable references.
Agent Team Go
Build and run AI-native agent teams in Go with custom skills, auto-install, and Feishu/Telegram gateways.
AI Native Capabilities
Spec-first agentic AI capability patterns for decision intelligence, commerce, engineering, supply chain, and compliance
Pulci
Quality gates that run between your AI agent's edits.
Cachly Kotlin
Official Kotlin SDK for cachly — persistent AI memory and managed Valkey/Redis with semantic caching. Free tier, EU-hosted.
Disco Bus
Push-based agent-to-agent message mesh with optional Discord mirroring. Tiny (~1000 LOC), push-not-poll, MCP-ready. Agents wake instantly on inbound; humans watch traffic in Disco…
Cachly Swift
Official Swift SDK for cachly — persistent AI memory and managed Valkey/Redis with semantic caching. Free tier, EU-hosted.
Veridge
The always-fresh, low-token map of a whole project — docs, code (to the symbol, multi-language), decisions — unified in one graph, ranked by PageRank, served to AI and humans. Zer…
Msw Skill
AI agent skill for MSW 2.0 — 20 rules: handler design, REST/GraphQL, test patterns. Works with Claude Code, Cursor, Codex, Windsurf.
Starter Skills
Free, MIT-licensed starter skills for a private, self-hosted AI assistant — back itself up, capture ideas, search your notes, self-reflect. Built to adapt, not just copy. The free…
NOpenCode
A .NET SDK that brings OpenCode's AI engine into your applications. Express your intent in natural language.
Agentlib
Agentic Knowledge Navigation — preprocessed books + a skill that teaches AI agents to navigate efficiently. No server, no vector DB, just files + instructions. Claude Code plugin.
Oci Kafka MCP Server
AI-native MCP server for OCI Kafka — 42 tools, enterprise security, AI diagnostics
Coracle
Personal-machine AI orchestrator: routes work between free-tier big AI (Gemini/Groq/Ollama Cloud + browser fallbacks) and local Ollama models (qwen2.5 + qwen2.5-coder) with a sing…
Agentic Architecture Audit
Evidence-first architecture audit for AI-assisted codebases. A discovery pass records what a repo claims to be. An eleven-phase audit tests those claims, with citations.
Dharma Tech
Buddhist AI — find texts, compare traditions, track practice, run Computational Madhyamaka. Local LLM + knowledge graph. No API keys.
Janus Assistant
Janus - a local personal AI assistant: runtime, CLI and Telegram bridge, powered by agentify-core. Installable via curl, npm or pipx.
Mini Browser
Token-efficient web search and browsing for AI agents. Search smarter, spend less tokens, get better answers.
Cachly Python
Official Python SDK for cachly — persistent AI memory and managed Valkey/Redis with semantic caching. Free tier, EU-hosted.
Bayesbelief
Privacy-preserving hierarchical Bayes for multi-tenant fleets: empirical-Bayes partial pooling, cold-start mitigation, zero-dependency TypeScript.
Client Intelligence Assistant
Production-ready AI platform for client intelligence powered by LangGraph, RAG, and MCP.
Deepseek Orchestrator
DeepSeek 三角色自动编排 dsh 插件：干活 v4-flash / 思考 v4-pro / 独立审查，规则路由自动分级，dsh 官方侧边栏面板 + CLI/MCP 多通道，零依赖
Kegg String MCP
LLM workflow to annotate genes with provenance and verification: MCP server exposing KEGG, STRING and PubMed as model-callable tools, plus an agentic annotation pipeline that veri…
Vibe Coding Wiki
178-term bilingual taxonomy for AI-assisted programming: vibe coding, agentic engineering, Claude Code, MCP, RAG, and more. Static HTML/JS, per-layer async load, offline cache, pr…
Skillfoundry
🔥 SkillFoundry (forge) — open, vendor-neutral toolchain for authoring, validating, testing & publishing AI agent skills. Works across Claude Code / Cursor / Codex / OpenCode / Cli…
Nanbeige4.2 3B Mps Fix
Deployment-blocking bugs, a Looped-Transformer memory bound, and a chunked-prefill fix for running Nanbeige4.2-3B on Apple Silicon (MPS)
Web Research
Self-hosted web research platform combining Firecrawl, SearXNG, Ollama and Playwright with unified REST and MCP APIs.
VisaRadar
Cross-references LLM-extracted job posting data against real DOL visa-sponsorship filing history — an evidence-backed sponsorship check, not another curated job board.
AI Keystone
"The INFRA Layer" that "Agentic AI applications" Run on. If the LLM is the CPU, this is the Operating System: Durable Agent Execution, A permissioned tool Boundary, Memory Tiers,…
PromptRack
Version prompts. Run them across models. Find what works.
IThome 2026ironman
iThome 2026 文章儲存庫 - 將AI導入企業日常，從想像、認知到設計、實作然後收尾。
Disposition
A personalization layer for AI-assisted coding. Learns how you write, then steers your AI coding tool to produce code that reads like you wrote it, so it stays maintainable by han…
Biomedical Rag Research Assistant
Biomedical RAG assistant for PubMed documents featuring hybrid ChromaDB and BM25 retrieval, MiniLM embeddings, RRF, grounded citations, guardrails, MCP tools, FastAPI, Streamlit,…
Autoproduct Design
Design docs for a full-lifecycle multi-agent software development system: opportunity sensing → market → PRD → spec → coding → review → test → deploy → evidence → kill decisions.…
Agentframe
OOP-native Python agents: state is fields, tools are methods, prompts are docstrings, contracts are types, actions are code.
Pi Agent Ha
Home Assistant add-on that runs the pi coding agent in a browser terminal inside your dashboard.
Bicameral
Two minds, one diff. Self-improving two-model AI coding agent for Claude Code: Claude plans and reviews, Codex writes the code, every step verified and learned from.
Agent Tool Interop
Research: which coding-agent CLIs expose a tool schema that a third-party inference backend can actually drive. Family A/B taxonomy, a drivability criterion, a compatibility matri…