"FIRST" matching MCP servers:

Long AI conversations fail in predictable ways. Context-First fixes all four: Failure Mode What Goes Wrong Context-First Solution Context Drift AI forgets earlier decisions and intent as the conversation grows context_loop + detect_drift continuously re-anchor every turn Silent Contradiction New inputs silently overrule established facts — the AI doesn't notice detect_conflicts compares every inp

Enables developers to enforce versioned data contracts for MCP tool schemas, with additive-only evolution checked in CI, SHA-256 pinning against rug pulls, and signed handles instead of protocol sessions.

Provides a single MCP tool that returns the current time for a requested timezone, connectable to Claude Code and Claude desktop over stdio or Streamable HTTP.

Ad-free web scraping and search exposed as 3 MCP tools. fetch_page, fetch_pages_batch, search_web. Works with Claude Desktop, Cursor, Cline.

Enables querying the Urban Institute's Education Data Portal — harmonized federal data on schools, school districts, and colleges — through a discovery-first set of tools that search and describe datasets, fetch raw records or aggregated statistics, decode coded values, and resolve institution names to IDs, with no API key required and read-only access to the live API.

A local-first, no-API-key MCP server that enables LLMs to search the web, fetch pages, and read documents using multiple engines and smart fallbacks.

Model Context Protocol server exposing the bankstatementparser library as first-class agent tools for reading bank statements.

Local-first memory for coding agents. Discovers the memory, instructions, and rules Claude Code, Codex, and Cursor already wrote on your machine, combines them into one canonical Markdown tree, and serves grounded, cited recall through tools like ask_memory and canonical_memory.

Embedded memory and retrieval engine for AI agents, providing local-first memory with MCP support for multi-agent access control.

Local-first crypto portfolio aggregation across exchanges (Bybit, Binance), EVM and Solana wallets, and Polymarket. Read-only credentials, runs as a local MCP server. Data aggregation only, not financial advice.

A local, read-only MCP server (plus CLI) over your own stock/ETF transaction log. A deterministic core computes drawdown-first risk with bootstrap confidence intervals and validates targets with walk-forward verdicts. The assistant narrates, but the number fence makes it structurally impossible for the model to produce a figure.

AgentShare delivers structured product search and pricing signals for AI agents over REST and MCP (Streamable HTTP). Responses include freshness & coverage metadata so agents can reason about data recency. API keys secure billed endpoints; public discovery at /agent.json and /mcp.json. Currently integrates connected marketplaces and affiliate feeds – roadmap expands to global e-commerce (AliExpre

A local-first multilingual dictionary, thesaurus, translation, lexical-relations, semantic-neighbour, and English wordplay server for the Model Context Protocol (MCP). It operates fully offline with read-only query access and no data mutation while running.

Privacy-first OSINT scanning MCP server that aggregates ~25 sources into a risk report, running locally with no data leaving your machine.

Enables AI assistants to control running Neovim sessions via RPC socket, supporting command execution, state inspection, and LSP actions. Automatically discovers Neovim instances and supports multi-instance management on Linux and macOS.

Agent-first cognitive substrate.

Enables transcription of videos and audio from 1000+ platforms (YouTube, Bilibili, TikTok, etc.) using subtitle extraction first, then local Whisper transcription, with support for long videos, async tasks, and Chinese ASR optimization.

DBeast MCP is a robust, security-first Model Context Protocol server designed to empower AI coding assistants (like Claude Desktop) with deep PostgreSQL database insights. Instead of just running basic SQL queries, DBeast equips your AI agent with tools to inspect active locks, audit slow queries, analyze table schemas, and safely generate index recommendations in real-time. Built specifically for

An MCP server that exposes a local notmuch email database to an LLM client such as Claude. It is read-first: searching, reading, and understanding mail is always available; writing anything (drafts, tags, exported files) requires an explicit opt-in flag and is confined to clearly bounded locations.

Enables text-first agents to inspect images via focused questions, returning compact, checked evidence packets. Reduces visual context by ~90% while preserving expected fields.