no code · no API key

How to Add Web Search & Research to Your AI Agent (MCP, No API Key)

Give your AI agent real research skills — web search, clean page reading, and cited dossiers — via an MCP server, with no external API key.

Open Deep Research MCP on Apify →

You want your AI agent to research the live web, but every search API wants its own key. Here's how to give Claude or Cursor web research with Deep Research MCP — the search runs on the platform, so there's no separate key.

Step by step

1

Start the MCP server on Apify

Open the Deep Research MCP actor and run it in Standby mode to get your MCP endpoint URL.

2

Connect your agent

Add the /mcp endpoint to Claude Desktop, Cursor, or any MCP client's config.

3

Ask it to research

Your agent now has web_search, read_url, and deep_research tools. Ask a question and it pulls live sources.

4

Get a cited dossier

deep_research returns sources with relevance scores, published dates, and query-relevant highlights — your agent writes the synthesis with [n] citations.

What you get

An AI agent that researches the live web and returns citation-ready answers — without you managing an OpenAI, Tavily, or Perplexity key.

💡 Automate it. It works with any MCP client — Claude Desktop, Cursor, Windsurf — and pairs with automation in n8n for scheduled research runs.

Try it free to start

Run Deep Research MCP on Apify — pay only for results, no subscription. New to Apify? It's free to start.

Open on Apify →

FAQ

Do I need an OpenAI or Tavily key?

No. The agent does the synthesis; the server does search + reading.

Which clients work?

Claude Desktop, Cursor, Windsurf, and any MCP client.

→ Full Deep Research MCP guide