MCP Server Directory

Read-only macro risk analysis, evidence, signals, and source-backed research from l0g.fr.

STDIO/SEE MCP Server for Apache Druid by iunera that provides extensive tools, resources, and prompts for managing and analyzing Druid clusters.

Chrome automation over raw CDP with multiple tabs, cached accessibility-tree refs, server-side multi-step flow execution and a Python scripting API.

Ad management across Google, Meta, LinkedIn, Microsoft, Reddit and TikTok: create campaigns, adjust budgets, generate creatives and pull performance data.

Policy-based governance for agent tool calls across LangChain, OpenAI, Anthropic and MCP: YAML policies, approval gates, risk assessment and audit logging.

Repo-native task continuity for coding agents: a reviewed task-state ledger of decisions, dead ends, evidence and checkpoints that later sessions resume from.

Governed MCP runtime for the Apple ecosystem, available on macOS for Notes, Mail, Calendar, Reminders, Shortcuts, and more.

Neuroscience-inspired local agent memory that stores episodes, consolidates knowledge over time and builds a personal knowledge graph with semantic recall.

EXPERIMENTAL MCP server for Google's Antigravity CLI (agy) — subscription-backed second opinions

Unified MCP server for multi-LLM consultation — routes Gemini, Codex, Claude, Grok, Ollama, and Antigravity behind runtime availability checks

Compiles TypeScript to validated Swift for Apple platforms: App Intents, SwiftUI views, WidgetKit widgets and full apps.

Bridge Codex and other MCP clients with Anthropic Claude Code CLI for read-only second opinions

Bridge Claude with OpenAI Codex CLI for AI-to-AI collaboration, code review, and second opinions

Returns active decision records for a path before an agent edits it: constraints, ruled-out alternatives and warnings, stored as git trailers and notes.

Self-improving tool discovery: searches registries, installs MCP servers as subprocesses without restarts, and persists tool knowledge across sessions.

Bridge Claude with Gemini CLI for AI-to-AI collaboration, code review, and second opinions

Bridge to GenieACS (TR-069 ACS) instances: device data, firmware management and CPE actions such as reboot, parameter refresh and firmware download.

Local code review for you and your coding agent: the agent opens a review of its changes, you comment on lines, it replies and resolves. Reads the git worktree directly, uncommitted work included; nothing leaves your machine.

Consult Grok through xAI's metered API with exact model selection and no fallback

A secure on-demand gateway for long-tail MCP servers with lazy calls and declarative Method Mode.

Human-governed memory in local, domain-separated vaults with hybrid retrieval; also reads other coding agents' output. Requires the Mnemosyne OS desktop app.

Shared community repository of insights, decisions, warnings and patterns with semantic search, a 3D interactive globe and automated content moderation.

Bridge Claude with local Ollama LLMs for private AI-to-AI collaboration — no API keys, fully local

Structured file edits for agents: JSON, YAML and TOML by selector, Markdown and AST edits, dry runs and multi-file transactional batches.

Compile browser automation into deterministic plans that run in your logged-in Chrome without runtime LLM calls, detecting drift when sites change.

Agent-first MCP server for WeChat Mini Program debugging, automation, and lightweight testing

One place for every AI agent's pages and docs: versioned links to share, search and update.

Local-first, read-only FinOps assistant: analyze AWS, Azure, GCP, Kubernetes, SaaS and AI bills, find waste and get fixes as pull requests.

Android development: build APKs, launch emulators, manage devices, automate UI and analyze logs.

Australian Football League teams, ladder standings, results, tips and power rankings via the Squiggle API.

Cross-session memory for AI agents - knowledge graph, scoring, semantic search

Token-efficient browser automation: semantic page snapshots and stable element IDs instead of raw DOM or screenshots, driving Chrome over CDP.

Governance proxy for MCP servers: allowlist, forced dry run, human confirmation, blast radius, audit

Query, analyze and monitor Azure resources at scale using Azure Resource Graph.

Let AI agents interact with real BLE hardware — scan, connect, read, write, and subscribe.

AI browser tools with Chrome CDP. Navigate, screenshot, interact. Multi-provider failover.

Cognitive browser automation that thinks like your users—and helps AI agents navigate too. Simulate real user cognition with abandonment detection, constitutional safety, chaos engineering, and UX friction discovery. Sites that pass CBrowser's cognitive t

Molecule names or SMILES to 2D structures, reactions, mechanisms, spectra - offline via RDKit.

Read Outlook email attachments via Microsoft Graph, parsing PDF, Word (including embedded images), Excel and text files into structured content.

MCP server for AI interaction with live Smalltalk images

Architecture docs and patterns for NestJS + Nuxt full-stack apps

Run code in multiple programming languages locally via Docker.

Local repo context for coding assistants with semantic search and graph relationships.

DeepSeek web vision model as an MCP server with browser auto-login.

Access Japanese financial disclosures (EDINET). Search companies, retrieve BS/PL/CF statements.

Manage and control ESP RainMaker devices.

FAOSTAT data for 245 countries: crops, trade, food security, and emissions via 23 MCP tools.

FHIR MCP server with SMART Backend Services, response compaction, FHIRPath, and guarded writes.