Last verified
n8n Research Agent
VerifiedAn n8n workflow agent that searches the web, summarizes sources, and emails findings
Agent details
- Kind
- n8n
- Autonomy
- supervised
Tools
- web_search
- http_request
- email_send
- claude_message
"A research digest in your inbox every morning, built in an afternoon."
What is the n8n Research Agent?
The n8n Research Agent is a pre-built automation workflow template for n8n that chains web search, AI summarization via Claude, deduplication, and email delivery into a repeatable agent pipeline. It runs on demand or on a schedule and requires no code, only configuration of API keys and search parameters.
How does it work?
The workflow has four main nodes: a trigger (manual or scheduled), a web search node that queries a configured topic, a Claude API node that summarizes each result, and an email node that sends the compiled digest. A deduplication node between search and summarize filters out URLs seen in previous runs, so each digest contains only new findings.
For teams that need multi-agent coordination where research is one step in a larger pipeline with decision-making and task execution, CrewAI provides a more flexible framework with agent-to-agent communication and role-based task assignment.
When should you use it?
Use the n8n Research Agent when you need a low-code, repeatable research workflow: monitoring a competitor, tracking a technology topic, or aggregating news about a market segment. It is the right fit when you want something running within an hour and have no need to write agent framework code.
Frequently asked questions
Does the n8n Research Agent require an n8n Cloud account or can I self-host?
It works with both n8n Cloud and self-hosted n8n (Docker or npm). The workflow template imports as a JSON file via the n8n UI. Self-hosting keeps all data on your infrastructure; n8n Cloud handles the infrastructure but stores workflow data on n8n's servers.
Which Claude model does the n8n Research Agent use for summarization?
The default template calls the Claude API via the n8n Claude node, configured to use claude-haiku for cost efficiency on high-frequency research runs. Switch the model in the node settings to claude-sonnet or claude-opus if summary quality matters more than cost.
Can the Research Agent monitor a topic on a schedule rather than on demand?
Yes. Replace the manual trigger node with a Schedule Trigger node and set a cron expression (e.g., daily at 7am). The agent will run on each tick, search for new content on your topic, and email a digest, functioning as a continuous monitoring feed.