n8n
Live web search inside an n8n workflow
An n8n workflow that calls a model stalls the moment the model needs a fact it was not trained on. Gyaansetu returns the passages that answer the question, scored, so the AI node can cite instead of guess.
By signing in you agree to the Terms and Privacy Policy.
Set it up
- 1
Make an API key
Sign in with Google at the dashboard and create a key. The free monthly allowance is 1,000 Credits and needs no card.
- 2
Add the search step
Use the Gyaansetu community node if it is installed, or an HTTP Request node: POST to /search with the key in the Authorization header.
- 3
Feed the passages to the model
Pass each result's content and url to the AI node. Advanced depth returns the paragraphs that answer the query rather than ten links.
HTTP Request node
Method: POST
URL: https://web-search.gyaansetu.com/search
Header: Authorization: Bearer {{ $credentials.gyaansetuApi.apiKey }}
Body (JSON):
{
"query": "{{ $json.question }}",
"search_depth": "advanced",
"max_results": 5
}The community node (n8n-nodes-gyaansetu) wraps the same two calls, so a workflow can switch between it and the HTTP node without changing the data shape.
Questions people actually ask
Try it before you decide
- 1,000 Credits free every month, no card
- Tavily-compatible requests, so an existing SDK works by changing the base URL
- MCP endpoint at /mcp for Claude Code and Cursor
- Failed requests cost nothing
By signing in you agree to the Terms and Privacy Policy.