21 servers. 17 providers. Sourced connection instructions.
Search the public web, retrieve page content and use optional browser tools through Bright Data’s hosted or local MCP integration.
Streamable HTTP · hostedBright Data API token in the token query parameter.
Local stdio · localAPI_TOKEN environment variable.
Inspect Evomi proxy products, targeting and usage, rotate sessions, and run scraping jobs through a local Python MCP server.
Local stdio · localEVOMI_PUBLIC_API_KEY environment variable; a separate scraper key can optionally override automatic lookup.
Choose keyless, OAuth or API-key access to Firecrawl search and extraction tools. Check which operations need an account before connecting your client.
Streamable HTTP · hostedNo key or account; daily per-IP request and credit caps apply.
Streamable HTTP · hostedOAuth sign-in in a compatible client; no API key pasted into this profile.
Streamable HTTP · hostedAuthorization bearer header; replace YOUR_FIRECRAWL_API_KEY in your private client configuration.
Search, extract pages, use browser tools and configure proxy connections through Novada’s hosted endpoint or local npm package.
Streamable HTTP · hostedAuthorization: Bearer YOUR_NOVADA_API_KEY header, or OAuth sign-in; Bearer is the only key method.
Local stdio · localNOVADA_API_KEY environment variable.
Connect an agent to Oxylabs scraping and AI Studio tools, with separate credentials for each product.
Streamable HTTP · hostedWeb Scraper API token in Authorization: Bearer (recommended) or X-Oxylabs-Api-Token. Alternatively, username and password via Authorization: Basic or the X-Oxylabs-Username/X-Oxylabs-Password headers.
Connect to Scrapfly’s hosted service for page retrieval, detailed scraping and browser-backed tasks.
Streamable HTTP · hostedOAuth sign-in to a Scrapfly account.
Capabilities and connection requirements come from the linked primary documentation. Review dates describe our evidence, not provider uptime.
Start with the task, then check the required account and connection profile. A provider can expose several capabilities while requiring different products or permissions for each.
Start with web search when you need source URLs before deciding which pages to read. Check the selected search tool and supported engines in the provider’s listing.
Documented examples: Bright Data MCP, Decodo MCP, Oxylabs MCP.
Compare 15 integrations →Use scraping and extraction when you already have URLs. Compare returned formats and the provider product required for those tools.
Documented examples: Bright Data MCP, Decodo MCP, Oxylabs MCP.
Compare 19 integrations →Check browser tools when your task involves navigation or page interactions. Some providers require a separate mode, product or tool group.
Documented examples: Bright Data MCP, Oxylabs MCP, Scrapfly MCP.
Compare 6 integrations →Look for specific credential, targeting or rotation tools. Review which calls expose credentials or change account state before enabling them in an agent.
Documented examples: NodeMaven MCP (community), Singapore Mobile Proxy MCP, Evomi MCP.
Compare 4 integrations →Use directory tools to compare provider records before buying or connecting. Working Proxy Sites exposes its reviewed provider and MCP catalogs through public reads.
Documented examples: Working Proxy Sites MCP.
Compare 1 integration →A hosted profile supplies the provider's MCP endpoint. Match its transport and authentication to your client: the reviewed templates may use API keys, account credentials or an OAuth sign-in.
Compare Streamable HTTP profiles →A local stdio profile starts a server package on your machine. Check its package version, runtime and environment variables before copying the client configuration.
Compare local stdio profiles →Every setup page links the documentation supporting its endpoint, client format and credentials. Confirm that profile's requirements before choosing a client.
This is a directory of provider integrations and their direct connection instructions. If you want to route several MCP servers through a gateway, that is a separate setup. Working Proxy Sites provides catalog data through its own MCP server; visitors connect directly to each listed provider for its tools.
Check the listed tools. A scraping or search tool does not establish proxy purchasing, credential management or rotation controls. Use the proxy-management filter for integrations with that documented capability, and read the limitations on the setup page.
Check the account and billing sections of the integration you choose. Server software, an MCP endpoint and the provider products used by its tools have separate requirements. Each listing links the source for its stated billing basis.
The evidence label identifies the level of review. “Documentation reviewed” means we checked the cited sources. “Connection tested” requires a dated test and a description of what succeeded. Review-due notices show when a source check needs refreshing.
Each listing has a documented MCP interface, a named publisher, usable setup instructions and primary sources. Capabilities reflect specific tools or documentation. Each record has a review due within 30 days; overdue records are marked visibly. A documentation review does not mean we connected to a paid service.
Publisher labels distinguish official integrations from reviewed community projects. Community entries identify the independent maintainer and installation source. Inclusion and alphabetical ordering do not depend on payment or a link back.
Use the JSON catalog or our MCP discovery tools to reuse these records. Cite the individual setup page or this comparison, and send corrections through our contact address.
