MCP servers

Servers that give an agent tools and data over the Model Context Protocol. Each entry says what the server exposes and links to its canonical source — Sartbot indexes, it does not host third-party projects.

Hosted first-party endpoints live under /mcp/<name>/. Long-form integration notes for those servers get their own page here when there is more to say than a card can hold.

Hosted by Sartbot

First-party servers, published and documented here.

  • Real-time quotes, fuzzy symbol search and historical K-lines for China A-share, Hong Kong, US and Taiwan markets — one Streamable HTTP endpoint, free upstream data, no API key.

    • stock
    • market data
    • quotes
    • k-line
    sartbot.com/mcp/mcp-stock-analyst

    Apify Standby · v0.2.2 · US · HK · CN · TW · Streamable HTTP

  • A curated, machine-readable catalogue of free tiers and low-cost token promotions for AI models and inference services — eight tools, four resources and three prompts, human-verified daily.

    • free tier
    • credits
    • promo
    • deals
    sartbot.com/mcp/free-and-cheap-tokens

    Apify Standby · v1.1 · 8 tools · 4 resources · 3 prompts · Streamable HTTP

  • Publishes a deterministic daily ethics feed for AI agents from a four-module curriculum — Goals, Principles, Case Studies, Conduct — so an agent keeps meeting the material without being asked. 10 read-only tools, 9 ethics:// resources, 1 prompt; pay-per-event, no monthly fee.

    • ai ethics
    • responsible ai
    • safety
    • alignment
    sartbot.com/mcp/ethics-for-ai-mcp

    Apify Standby · 10 tools · 9 resources · 1 prompt

Widely used servers

Reference and community servers worth knowing before you build your own.

Search all MCP servers →
  • Reference server giving an agent controlled read and write access to a sandboxed set of local directories.

    • files
    • filesystem
    • read
    • write
    github.com/modelcontextprotocol/servers

    Official reference · @modelcontextprotocol/server-filesystem

  • Read repositories, triage issues and pull requests, and review code through the GitHub API without leaving the agent.

    • github
    • git
    • issues
    • pull requests
    github.com/modelcontextprotocol/servers

    Official reference · @modelcontextprotocol/server-github

  • Persistent knowledge-graph memory so an agent can remember entities and relations across sessions.

    • memory
    • knowledge graph
    • context
    • persistence
    github.com/modelcontextprotocol/servers

    Official reference · @modelcontextprotocol/server-memory

  • Drive a real browser through Playwright — navigate, click, fill forms and capture screenshots, using the accessibility tree instead of screenshots where possible.

    • browser
    • automation
    • playwright
    • testing
    github.com/microsoft/playwright-mcp

    Microsoft · @playwright/mcp

  • Pulls current, version-specific documentation and code examples for a library straight into the model's context.

    • documentation
    • docs
    • library
    • api reference
    github.com/upstash/context7

    Upstash · @upstash/context7-mcp

  • The protocol itself: transports, tools, prompts, resources and sampling. Start here before writing a server.

    • protocol
    • spec
    • documentation
    • transports
    modelcontextprotocol.io

    Specification · Model Context Protocol