Fetch MCP
Fetch any URL or web content from AI
How to Install Fetch MCP
$npx @modelcontextprotocol/server-fetchRequires Claude Desktop, Cursor, Windsurf, or another MCP-compatible client.
About Fetch MCP
A general-purpose URL fetcher for AI — pulls and converts pages to markdown for the model.
Fetch MCP is a APIs & Data MCP server designed for engineers and ops teams. It enables connecting AI assistants to third-party APIs and data feeds by wrapping HTTP endpoints with auth and schema as MCP tools. Key capabilities include auth helpers for common providers, pagination and rate-limit handling, endpoint allow-lists, and structured tool calls that work cleanly inside an LLM context. It integrates with Salesforce, HubSpot, Mixpanel, Stripe, and most REST/GraphQL APIs, and is best suited for engineers and ops teams who need internal admin apis.
Key Features
- http
- fetch
- web
- Auth helpers for common providers
- Pagination and rate-limit handling
- Endpoint allow-lists
Pricing
- Core MCP server
- Community support
- Works with any MCP client
- Everything in Free
- Higher usage limits
- Priority support
- Everything in Pro
- SLA & SSO
- Dedicated support
Tier details are indicative — visit the Fetch MCP website for current pricing.
Pros & Cons
Pros
- Fills the long tail when no purpose-built server exists
- Reuses your existing API surface
- Composable with other MCP servers
Cons
- Quality depends on upstream API design
- Rate limits and quotas need tracking
- Auth refresh logic can be fiddly
Best For
- Internal admin APIs
- CRM and analytics queries
- Glue between specialised MCP servers
- Replacing one-off webhook scripts
Screenshots
Screenshots coming soon —
Submit yours →Our Take on Fetch MCP
Boring, essential. Every agent needs a way to read a URL.
Alternatives to Fetch MCP
Real-time web search for Claude and AI models
See how Fetch MCP compares to Brave Search MCP →Browser automation and web scraping for AI
See how Fetch MCP compares to Puppeteer MCP →