TLDR: The n8n You.com node now covers the full You.com product line, growing from 3 operations to 7. Three are new: Answer, Finance Research, and Get Research Task. Research adds a background mode for long-running tasks, so you can queue a job and poll its task id instead of holding a connection open, along with a new frontier effort level, domain-level source control, and JSON Schema output. Web Search adds include, exclude, and boost domain filters plus a new Extraction collection. Contents accepts multiple URLs and a max_age freshness threshold. All You.com nodes are now available as tools for the n8n AI Agent node. One thing to check before you upgrade: livecrawl and livecrawl_formats on Web Search have been replaced by Extraction.
Synchronous research calls have always worked from n8n. The trouble starts when a request runs long enough to outlive the execution window, which an exhaustive or frontier query will do. Getting around that meant building the polling loop by hand: an HTTP Request node to kick the job off, a Wait node, an IF branch checking a status string you had to go find in the docs, and a loop back around.
That pattern now ships as two operations. Along with it: Answer, Finance Research, and enough new parameters on the existing operations that most working workflows deserve a second look.
Three New Operations in the n8n You.com Node
Answer (POST /v1/answer) returns a synthesized answer with citations drawn from web search results. It takes the same filters as Web Search: freshness, country, language, safesearch, and domain filters. Reach for it when the next step in your workflow needs prose and sources rather than a ranked list of URLs.
Finance Research (POST /v1/finance_research) returns finance-grade answers with citations at two effort levels, deep and exhaustive.
Get Research Task (GET /v1/research/{task_id}) reads the status of a research task you queued in the background. It's the second half of the pattern below.
Background Mode Solves the Long-Running Research Problem
Research has a new background mode. Set it and the node returns a task id right away instead of blocking. Pass that id to Get Research Task, put a Wait node in front of it, and loop.
Loop on the status, not on the clock. A three-minute Wait works fine until the day a query takes four, and then your workflow hands the next node a result that doesn't exist yet. Keep polling until the returned status is "completed".
Research effort now has five levels, and the highest is frontier. Frontier runs in background mode only.
The n8n You.com Node Runs as an AI Agent Tool
All You.com operations are available as tool sub-nodes for the n8n AI Agent node. Connect one to an agent, and the agent decides on its own whether a given question needs a web search, a page fetch, or a full research task. You no longer have to guess at build time which branch will need a lookup and wire a search step into that spot.
New Parameters on Search, Research, and Contents
Web Search gets three domain filters: include_domains, exclude_domains, and boost_domains. Each accepts up to 500 values, and include_domains is mutually exclusive with the other two. There's also a new crawl_timeout, accepting 1 to 60 seconds with a default of 10.
Research gets a source_control collection covering domain filters, freshness, and country, so a research run can be pointed at a source set you define rather than the open web. It also gets output_schema: supply a JSON Schema and the result comes back structured. That matters when the node after this one is a database insert instead of a Slack message.
Contents now takes multiple URLs through the + button, with CSV strings still accepted as a fallback for anything already passing a joined expression. The new max_age parameter sets a freshness threshold, and cached content older than that gets re-fetched.
Livecrawl Is Gone. Migrate to Extraction.
This is the change that will break a workflow that currently works. Web Search no longer accepts livecrawl or livecrawl_formats.
The replacement is the Extraction collection. Set Extraction Mode to highlights or full_page. For full page content, open the Full Page sub-collection and choose your extraction_formats: markdown, html, or both.
The You.com node shows up in the n8n node search panel as soon as you type "you.com." Grab an API key and run Answer against a question you already know the answer to. That's the fastest way to see the citation format before you build anything on top of it.
