Browse the catalog

Adaptive MCP memory system for AI applications. Learns which retrieval strategies work for your data, scores results using cognitive science models, builds a knowledge graph automatically, and validates every parameter change against real query history before adopting it. Patent pending.

Agent budget metering prepaid credit wallet & velocity caps (Skyfire style)

Sovereign, MIT-licensed MCP server for professional service workflow, wrapping the FallRecruit US SDK, enabling offline-capable, Ed25519-signed listing and transaction operations.

A Model Context Protocol server that enables Claude to directly interact with GitHub, including repos, files, branches, issues, pull requests, Actions, and code search using a personal access token.

Connect 2,000+ APIs with managed OAuth via Pipedream.

just a test of glama

Hosted Google Merchant Center MCP server: 44 actions (22 read, 22 write) on Google's Merchant API, switched on or off per account, with deletes off by default. Works with Claude, ChatGPT, Cursor and Claude Code.

Provides AI assistants with access to Tushare financial data API for querying Chinese stock market data, financial statements, and macroeconomic indicators.

Engineering memory for Claude Code — prevents re-investigating solved problems and repeating rejected architectural decisions across sessions and projects.

Exposes SymPy's symbolic mathematics functionality for algebraic simplification, calculus operations, equation solving, linear algebra, number theory, and special functions. Provides tools for differentiation, integration, limits, series expansion, and set theory operations.

Analyzes ambiguous Korean chat messages by interpreting nuance, emotional tone, and context, and suggests appropriate replies for social situations like apology, refusal, or setting boundaries.

An end-to-end example demonstrating how to use Model Context Protocol (MCP) over STDIO in .NET to expose server-side tools and consume them from a client, then feed tool results into a local Ollama-hosted LLM for reasoning and response generation.

This MCP server provides AI assistants with access to the atPlatform ecosystem for decentralized data sharing and storage, built using Dart with the at_client SDK and dart_mcp library for protocol integration. The implementation offers two core tools for listing atKeys with filtering by shared relationships and regex patterns, and retrieving atKey values with binary data detection, plus serves comprehensive atPlatform documentation as markdown resources covering protocol specifications, SDK usage, and authentication workflows. Built with custom logging integration, configurable root domain/port settings, and local atSign key management through the at_onboarding_cli, it serves developers building atPlatform applications who need conversational access to atSign data operations, teams exploring decentralized identity solutions, and AI assistants requiring detailed atPlatform documentation to guide users through implementation and troubleshooting of atSign-based applications.

Provides location information from coordinates or IP addresses using IPInfo.io, with a health check endpoint.

Connect Windsurf to live trend data: Google, YouTube, TikTok, Reddit, Amazon, news, Steam, npm and more. One MCP, free tier, no scraping.

Ads MCP - manage Google, Meta, TikTok, LinkedIn, Microsoft, Reddit & X Ads campaigns from Claude, ChatGPT, Cursor & any MCP client.

Creative AI Group delivers expert AI consultancy, custom AI tools, and MCP server implementations to help businesses automate, optimize, and scale. From SKILL.md documentation to production-ready AI systems — we build what moves the needle.

Enables AI to directly call FreeCAD for 3D modeling via the MCP protocol, providing 39 modeling tools including primitive creation, boolean operations, transformations, and export.

An MCP server that lets an AI agent operate a Logseq (DB) graph through safe, high-level tools for reading and managing notes, tasks, and projects without needing to know Logseq's internals.

This MCP server, developed by dwrtz, provides stateful, TUI-compatible terminal sessions for AI assistants. Built with Go and leveraging the mcp-go library, it offers two main tools: 'run' for executing commands in persistent bash sessions, and 'runScreen' for interactive TUI applications. The implementation focuses on maintaining session state, handling control sequences, and supporting multi-line outputs. It's particularly useful for AI systems that need to interact with command-line interfaces, enabling use cases such as file system navigation, text editing in vim, or running interactive programs like Python REPL without losing context between commands.