A web crawler and scraping API that records what changed between crawls.
Turn any site into clean markdown, crawl it on a schedule, and get the pages that were added, changed or removed — in your database, your vector store, or your AI assistant.
Website · Docs · MCP server · Pricing · Start free
- Scrape a page to markdown, text, HTML, links, structured fields or a screenshot.
- Crawl a whole site, or map what its sitemaps declare before fetching anything.
- Search the web and read each result's page in the same call.
- Watch a site over time: scheduled runs and a change record — pages added, removed and modified, field by field.
- Deliver to webhooks, Slack and other chat apps, or straight into 18 destinations: Postgres, BigQuery, Snowflake, S3, Pinecone, Qdrant and more.
You pay only for pages that were read: blocked pages, 404s and unchanged pages cost nothing.
The hosted MCP server needs no API key — you sign in and approve once.
claude mcp add --transport http mesharc https://mcp.mesharc.dev/mcpWorks with Claude Code, Claude Desktop, Cursor, Windsurf and any client that speaks MCP over HTTP. Setup for each client →
pip install mesharc # Python 3.10+
npm install mesharc # Node 20+from mesharc import MeshArc
arc = MeshArc() # reads MESHARC_API_KEY
page = arc.scrape("https://example.com/pricing")
print(page["markdown"])| Repository | What it is |
|---|---|
| mesharc-python | Python client for the MeshArc API, with the MCP server |
| mesharc-node | Node client for the MeshArc API, with TypeScript types |
| mesharc-mcp | The MCP server, packaged for directories and self-run setups |
All open-source code here is MIT-licensed.
Questions and feature requests: open an issue on the relevant repository, or write to hello@mesharc.dev. Security reports: security@mesharc.dev — see mesharc.dev/legal/security.