Get fast web research inside chat. An AI agent reads your message, chooses Google, Bing, or Yandex, and returns links with short summaries. This is helpful for marketing teams that need quick market checks and simple alerts.

The flow starts on a new chat message or from a manual test. Google Gemini interprets the request and uses a small memory window to keep short context. The agent calls Bright Data MCP tools to search across engines and formats the results in markdown. It then sends the final answer to a webhook so you can log or forward the response to other systems.

Self hosted n8n is needed because the MCP Client is a community node. You will need a Google Gemini API key, access to a Bright Data MCP server, and a webhook endpoint such as Webhook site. Expect faster research cycles, fewer manual checks, and a steady output you can use for SEO tasks, competitor tracking, and content planning.

• Chat message trigger starts research when a user sends a prompt
• AI Agent uses Google Gemini to understand the question and plan the search
• Simple Memory keeps a short window of context for follow up questions
• Bright Data MCP search tools run Google, Bing, and Yandex queries
• Results come back in markdown with URL, title, and description
• HTTP Request node posts the final answer to a webhook for notifications
• Manual test path seeds a sample query to verify the tool setup
• MCP Client lists available tools so the agent can pick the right one
• Dynamic engine selection lets the agent switch between search engines

• Reduce manual web checks from 30 minutes to 2 minutes per query
• Automate up to 80 percent of repeated search tasks
• Improve research consistency with one clear format of links and summaries
• Send results to any system using a single webhook URL
• Handle more volume by running Google, Bing, and Yandex from one chat

1. Import the template into n8n: Create a new workflow in n8n > Click the three dots menu > Select 'Import from File' > Choose the downloaded JSON file.
2. You'll need accounts with Google Gemini, Bright Data MCP and Webhook.site. See the Tools Required section above for links to create accounts with these services.
3. Use self hosted n8n because the MCP Client is a community node. Make sure the community nodes feature is enabled and that you can run external MCP tools.
4. Set up Google Gemini credentials: double click the Google Gemini Chat Model node, choose Create new credential, and follow the on screen steps. Create an API key in your Google Gemini account, then paste it into the credential. Give the credential a clear name.
5. Prepare Bright Data MCP: start or connect to your Bright Data MCP server as described in its docs. In the MCP Client nodes, open the Credential to connect with dropdown, click Create new credential for MCP Client STDIO, and follow the on screen steps.
6. Validate the MCP tools: run the MCP Client list all tools for Bright Data node and confirm that a tool named search_engine appears in the output.
7. Check the AI Agent configuration: confirm the Google Gemini model and Simple Memory are connected to the agent. Make sure the Google, Bing, Yandex search tools and the HTTP Request tool are added as available tools.
8. Configure the webhook: open the HTTP Request for Webhook Notification node and replace the sample Webhook.site URL with your own endpoint. Keep method as POST and include the chat_response field in the body.
9. Run a manual test: execute the When clicking Test workflow node, then run the next nodes to confirm the search tool returns results for the sample query. Review the node output for URLs, titles, and descriptions.
10. Test the chat flow: open the n8n chat interface for this workflow or use the provided chat URL. Ask a search question and confirm you receive formatted results and that your webhook receives a payload.
11. Troubleshoot failures: if no results appear, check that the MCP server is running, the engine parameter is set correctly, and that your Gemini API key is valid. Verify that each node has the correct credential selected.
12. Harden for production: replace Webhook.site with a secure endpoint you control and limit access to credentials using n8n credential permissions.