Repository navigation
Add MrScraper MCP server to partner registry - #265
MrScraper (ai-mrscraper) wants to merge 1 commit into
Conversation
Register the hosted MrScraper MCP server (https://mcp.mrscraper.com/mcp, Streamable HTTP) with its scraperAPI, webSearch and status tools, MIT license, docs and support contact. It declares two security schemes: a bearer API key (mrscraperBearerAuth), which is what Foundry uses, and OAuth 2.1 with PKCE (mrscraperOAuth) for MCP clients that support dynamic client registration or client ID metadata documents. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
|
MrScraper (@ai-mrscraper) please read the following Contributor License Agreement(CLA). If you agree with the CLA, please reply with the following information.
Contributor License AgreementContribution License AgreementThis Contribution License Agreement (“Agreement”) is agreed to by the party signing below (“You”),
|
|
Linda Li (@lindazqli) Yongjia Li (@yongjiaaaa) could you take a look when you have a moment? This adds the MrScraper MCP partner entry ( |
This PR adds one new partner entry,
partners/servers/mrscraper-mcp-server.json, for the hosted MrScraper MCP server. It is for the Microsoft Foundry tool catalog and the Azure API Center MCP registry. No other files change.MrScraper will also submit the Azure API Center MCP Registry Sign-up form with a link to this PR.
What MrScraper MCP does
MrScraper is a web scraping service. The MCP server gives AI agents live web data through three read-only tools:
scraperAPIwebSearchstatusAll three tools are annotated
readOnlyHint: trueanddestructiveHint: false.scraperAPIandwebSearchare alsoopenWorldHint: true. Each returns{status_code, data, headers, error?}, with credentials and cookies removed from the headers. Tool calls send the requested URLs, search queries and any options to MrScraper. Usage is counted against the account's MrScraper tokens.Endpoint and authentication
https://mcp.mrscraper.com/mcpremoteType: streamable-http)mrscraperBearerAuthhttp,scheme: bearerAuthorization: Bearer <MrScraper API key>. Keys are created at https://app.mrscraper.com/api-tokensmrscraperOAuthoauth2,authorizationCodehttps://api.app.mrscraper.com/oauth/authorize, token and refresh URLhttps://api.app.mrscraper.com/oauth/token, scopesscrape:read(scraperAPI, webSearch),account:read(status) andoffline_accessFor Foundry, use the API key. MrScraper's OAuth clients come from dynamic client registration or client ID metadata documents. The sign-up form says Foundry does not support dynamic client registration yet, so key-based authentication (credential name
Authorization, valueBearer <key>) is the method that works in Foundry today. The OAuth scheme is there for other MCP clients that support those registration methods. The endpoints come from the authorization server metadata athttps://api.app.mrscraper.com/.well-known/oauth-authorization-server. The protected-resource metadata is athttps://mcp.mrscraper.com/.well-known/oauth-protected-resource/mcp. NoauthSchemasvalue is set, because MrScraper has no managed OAuth client for Foundry yet. If you would like managed OAuth for Foundry, please let us know and we will look at registering a client for it.The entry
namemrscraper-mcptitlekind/vendormcp/Partnerremote/remoteTypehttps://mcp.mrscraper.com/mcp/streamable-httpcategoriesWeblicenserepositoryexternalDocumentationiconsupportContactInfouseCasessecuritySchemesmrscraperBearerAuth(bearer API key),mrscraperOAuth(OAuth 2.1 with PKCE)versionName/x-ms-previeworiginal/false(generally available)The layout follows the existing remote web-data entries (
brightdata-mcp-server.json,browserbase-mcp.json,parallel-search-mcp-server.json,exa-mcp-server.json). Declaring a bearer scheme and anoauth2scheme side by side followsgithub-mcp-server.json,neon-mcp-server.jsonandsupabase-mcp-server.json.summaryis one line of 134 characters,descriptionis 963 characters and the OAuth scheme description is 184 characters.Validation
I ran the three steps of
.github/workflows/validate-servers.ymllocally with Python 3.12 and check-jsonschema 0.38.2:Endpoint and link checks (no real credentials used):
GET https://mcp.mrscraper.com/mcpwith no credentials returns401, withWWW-Authenticate: Bearer error="invalid_token", error_description="Missing Authorization header", resource_metadata="https://mcp.mrscraper.com/.well-known/oauth-protected-resource/mcp".initializePOST with a made-up bearer value returns401with"error_description":"Invalid MrScraper API token", so the server checks bearer values as MrScraper API keys.https://api.app.mrscraper.com/.well-known/oauth-authorization-serverboth return200with JSON. The authorization server metadata lists the authorize and token URLs used inmrscraperOAuth, PKCES256, and the scopesscrape:read,scrape:write,account:readandoffline_access.200.Not covered here: authenticated tool calls, an OAuth sign-in, and an end-to-end connection from Foundry Agent Service. These need a MrScraper API key or account, which this submission environment does not have. The Foundry check (at least five queries through a custom MCP tool, as the sign-up form asks) will be done by the MrScraper team with an Azure subscription and reported in the form.
Contact
MrScraper Support, support@mrscraper.com. Docs: https://docs.mrscraper.com. We're happy to change anything the review calls for.
This PR comes from MrScraper, which runs this service.
🤖 Generated with Claude Code