OpenUnfurl
Zero-signup link-preview API that returns clean metadata (title, description, image, etc.) from any public URL. No API key required.
README
OpenUnfurl
A zero-signup link-unfurl / link-preview API. No account, no API key — one GET request.
Give it any public URL, get back clean preview metadata (title, description, image, siteName, favicon, Open Graph / Twitter Card, oembed) as JSON. Built so autonomous AI agents can use it as a tool — an agent has no human to do a signup.
Live: https://openunfurl.vercel.app
Quick start
curl "https://openunfurl.vercel.app/api/unfurl?url=https://github.com"
Sample response:
{
"url": "https://github.com",
"resolvedUrl": "https://github.com/",
"title": "GitHub · Build and ship software on a single, collaborative platform",
"description": "Join the world's most widely adopted, AI-powered developer platform where millions of developers, businesses, and the largest open source community build software that advances humanity.",
"image": "https://github.githubassets.com/assets/campaign-social-031d6161fa10.png",
"siteName": "GitHub",
"type": "object",
"favicon": "https://github.githubassets.com/favicons/favicon.svg",
"oembed": null,
"fetchedAt": "2026-05-17T00:00:00.000Z",
"engine": "static-html-v0.1",
"note": "v0.1 parses static HTML only — no JS/SPA render"
}
Any missing field is null. Errors return JSON with an error key and an
appropriate HTTP status (400 bad/blocked/missing URL, 422 fetch failed,
429 rate limited, 405 wrong method).
JavaScript
const base = "https://openunfurl.vercel.app";
const r = await fetch(
base + "/api/unfurl?url=" + encodeURIComponent("https://example.com")
);
const meta = await r.json(); // { title, description, image, favicon, ... }
Use as an MCP tool (AI agents / Claude / Cursor / LLM clients)
OpenUnfurl is also a remote MCP server, so an agent can call it as a tool with no signup, no API key, no OAuth.
- Endpoint:
https://openunfurl.vercel.app/api/mcp - Transport: Streamable HTTP, stateless (POST JSON-RPC 2.0, single
application/jsonresponse — no sessions, no SSE) - Exposes exactly one tool:
unfurl— input{ "url": "https://example.com" }, returns the same preview JSON as the REST endpoint (also asstructuredContent)
Drop this into any MCP client config (Claude Desktop, Cursor, or any client that speaks the Streamable HTTP transport):
{ "mcpServers": { "openunfurl": { "url": "https://openunfurl.vercel.app/api/mcp" } } }
Quick smoke test:
curl -s -X POST https://openunfurl.vercel.app/api/mcp \
-H 'Content-Type: application/json' \
-d '{"jsonrpc":"2.0","id":1,"method":"tools/call","params":{"name":"unfurl","arguments":{"url":"https://example.com"}}}'
Over MCP, an upstream error comes back as a tool result with isError: true.
Why
Managed link-preview alternatives — Microlink, OpenGraph.io, LinkPreview.net,
Unfurl.io, LinkPeek — all gate even their free tier behind a signup and/or an
API key. OpenUnfurl does not: it's a single anonymous GET (or one MCP
tools/call). That's especially useful to autonomous agents, which have no
human in the loop to complete a signup or paste in a key.
Limitations
- v0.1 parses static HTML only — no headless browser, no JS-rendered SPAs.
If a site renders its
<meta>tags client-side you get the static fallback. - Best-effort per-instance IP rate limit — a soft abuse brake, not a guarantee (serverless instances are ephemeral and not shared).
- SSRF-guarded — rejects localhost / private / reserved IP ranges.
- MIT licensed and self-hostable — see below.
Self-host — it's zero-dependency files
The entire API is two single zero-dependency Node serverless functions:
api/unfurl.js (REST) and api/mcp.js (remote
MCP). No npm install, no cheerio/jsdom. Deploy the folder to Vercel
(zero-config /api detection) or drop the handlers into any Node serverless
runtime.
git clone https://github.com/SolvoHQ/openunfurl
cd openunfurl
npx vercel --prod
License
MIT — see LICENSE.
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.