Turn a short chat prompt into fast research. The system searches the web, finds the strongest source, and returns a clear summary. Ideal for content teams, marketers, and sales researchers who need quick insights without manual reading.

Here is how it runs end to end. A chat window collects the topic. The flow sends that topic to Tavily Search with your API key. Results are filtered to only keep high confidence items above a 90 percent score, then the top link is selected. Tavily Extract pulls the full page text from that top link. The OpenAI Chat Model with a summarization chain condenses the content into a brief that is easy to read. Optional image fields can be included from search if needed.

Setup is simple but requires two credentials. Add your Tavily API key and connect your OpenAI account. Adjust the relevance threshold and any time range or domain rules in the search request. Teams often cut research from about one hour to a few minutes per topic. Use this for blog briefs, competitor snapshots, PR monitoring, or sales call prep. Once saved, open the chat window in n8n, type your topic, and get the summary back in seconds.

• Chat window trigger collects the search topic from a user
• Tavily Search runs a web query with depth, topic, time range, and image options
• Relevance threshold keeps only results above 90 percent confidence
• Top link is selected and passed forward for deeper processing
• Tavily Extract retrieves the full page text for the chosen URL
• OpenAI Chat Model and a summarization chain produce a clean, readable brief

• Reduce topic research from 60 minutes to 5 minutes
• Automate up to 80% of manual reading and copy pasting
• Improve relevance by filtering to 90% confidence scores
• Connect Tavily and OpenAI in one repeatable flow
• Handle 10 times more research requests per day

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 OpenAI and Tavily. See the Tools Required section above for links to create accounts with these services.
3. In your Tavily account, create an API key from the API page. Copy the key for use in n8n.
4. In n8n, open the node named Tavily API Key and replace any placeholder with your real Tavily key. Also open the Tavily Search Topic and Tavily Extract Top Search nodes and ensure the body includes your key using the api_key field.
5. In the n8n credentials manager, create a new OpenAI credential. Paste your OpenAI API key from your OpenAI account. Name the credential clearly and save it.
6. Open the OpenAI Chat Model node and choose your saved OpenAI credential under the Credential to connect with dropdown. Pick a suitable model and set max tokens if needed.
7. Open the Tavily Search Topic node. Map the query field to the chat input from the Provide search topic via Chat window node. Adjust search_depth, topic, time_range, and max_results as needed.
8. Review the Filter > 90% and Get Top Result nodes. Confirm only items with a score above 0.9 move forward and that the first ranked item is selected.
9. Open the Tavily Extract Top Search node. Ensure the urls field pulls the selected link from the prior node and that the request uses application/json.
10. Open the Summarize Web Page Content node. Review the prompt and summary length. Adjust tone or format to match your team style.
11. Enable the workflow. Open the n8n Chat window for the Provide search topic via Chat window node, enter a topic, and run a test. Confirm you receive a clear summary.
12. Troubleshoot common issues: a 401 error means the Tavily key is missing or invalid, empty results often need a broader query or time_range, and OpenAI errors may be rate limits. Adjust settings and try again.