You.com

Ground AI agents in live web data with You.com. Agents run web searches, fetch page contents, launch and poll deep research tasks, and pull finance-focused research to answer questions with current sources.

6 actions API key available

You.com website ↗ Platform Documentation ↗ Capability data verified 2026-07-27

A new vendor above the spend threshold enters the procurement queue

Web Search sweeps current news and filings mentions for the vendor, freshness-filtered to the last year

Get Contents pulls the full text of the most relevant pages as clean markdown

Start Research Task launches a Frontier-effort background job on the vendor's financial stability, litigation history, and customer reputation

Get Research Task polls until the synthesized report completes, with inline citations and the full source list

The procurement lead reads the report and spot-checks its citations before payment terms are offered

The decision, the report, and its sources file to the vendor record for the audit trail

What This Integration Enables

You.com treats research as an API. Not search, research: given a hard question, its agentic Research runs multiple searches, reads the pages, cross-references sources, and returns a synthesized Markdown answer with inline citations. The distinction matters for how FlowRunner uses it. An LLM without retrieval answers from memory; a search call returns links someone still has to read. You.com closes that gap, and FlowRunner agents wire its output into the workflows where current, sourced answers change what happens next. - Web Search returns LLM-ready results with freshness, country, language, and domain controls, plus optional live crawling - Get Contents converts batches of URLs into clean markdown for grounding, summarization, or RAG ingestion - Run Research handles synchronous questions; Start Research Task and Get Research Task run the deepest Frontier efforts in the background - Run Finance Research applies a specialized agent to companies, tickers, earnings, and macro questions - Structured JSON output schemas let research results feed downstream steps as data, not prose - Findings pause for [human-in-the-loop](/concepts/human-in-the-loop) review before they become decisions with money attached

Without FlowRunner

Research depth depends on the calendar Diligence gets whatever hours were left, so some vendors get an afternoon and some get a glance

Answers come from model memory LLM steps respond from training data, confidently describing a company as it existed two years ago

Findings arrive without provenance A summary lands in the deck and nobody can say which source, if any, supported the key claim

With FlowRunner

Every case gets the same sweep The search, read, and cross-reference loop runs identically for vendor one and vendor one hundred

Answers cite the live web Reports arrive with inline citations and source lists, freshness-filtered to the window that matters

Judgment stays with a named owner The agent researches; a person reads the citations before the finding becomes a decision

Use Case Scenarios

Account briefs that are current on the morning of the call

The evening before each customer meeting on [Google Calendar](/integrations/google-calendar), the agent runs Web Search on the account with a seven-day freshness window, pulls the substantive hits through Get Contents, and assembles a one-page brief: leadership changes, product news, funding, anything that moves the conversation. The brief posts to [Slack](/integrations/slack) before the rep's first coffee. Meetings start from this week's facts instead of last quarter's CRM notes.

A knowledge base that ingests the web with its sources attached

The team's retrieval pipeline needs current industry content, not just internal documents. The agent runs scheduled Web Search sweeps on tracked topics with domain boost lists favoring primary sources, fetches full text with Get Contents, and upserts the embedded chunks into [Pinecone](/integrations/pinecone) with source URLs in the metadata. Downstream answer flows cite their way back to the original page, and stale entries age out on the same schedule that brought them in.

Market context for the quarterly plan, researched to a deadline

Two weeks before planning, the agent launches Start Research Task jobs on each strategic question, demand signals in core segments, competitor moves, regulatory changes, and runs Run Finance Research on the public companies that anchor each market. Completed reports land in [Google Docs](/integrations/google-docs) with citations intact, and the strategy lead reviews each before it enters the planning pack. The team debates conclusions in the room, not search results.

Human-in-Loop Highlight

Run Finance Research is where this connector's output gets closest to money, and it is gated accordingly. A synthesized analysis of a counterparty's financial condition reads with the confidence of an analyst's memo, but it was assembled by an agent from public sources at machine speed, and the decision it feeds, extending six-figure credit terms, signing a multi-year vendor commitment, cannot be unsigned. So the diligence flow never lets the report flow straight into the decision record. It delivers the analysis with its inline citations and asks the owner to check the claims that carry the weight: "Report flags declining receivables quality at the target, cited to two sources. Confirm before terms go out?" The same discipline governs Frontier-effort Start Research Task jobs, which are the connector's slowest and most expensive calls: launched deliberately, in batches a person approved, not fired speculatively by a loop. Research compounds when it is trusted, and it stays trusted because a person signs what it concludes.

Agent processes routinely

Detects exception requiring judgment

Clear match Continues automatically

Ambiguous Routes to human via preferred channel

Human decides

Agent resumes with decision

Agent Capabilities

6 actions

Search

Web Search Runs a real-time web search returning LLM-ready results with title, URL, and snippets, plus related news, in one response. Supports freshness and date filtering, country and language targeting, safe search, domain include, exclude, and boost lists, and optional live crawling of full page content.
Get Contents Fetches and extracts the full content of specific pages by URL, returning clean markdown or HTML with title and metadata, up to a batch of URLs per call. The step that turns search hits into text a pipeline can use.

Research

Run Research Runs agentic Research synchronously: multiple searches, page reads, and cross-referencing, returning a synthesized Markdown answer with inline citations and sources. Supports freshness constraints and structured JSON output via an output schema. Deeper effort levels take longer; for Frontier effort, use the background pair below.
Start Research Task Starts a long-running Research job in background mode, returning a task_id and status immediately. The path for Frontier-effort research and anything too slow to run inline.
Get Research Task Polls a background Research task by task_id until status reads completed, then returns the synthesized answer, its citations, and the source list.
Run Finance Research Runs the specialized Finance Research agent on companies, tickers, earnings, and macro questions, returning a cited Markdown analysis. Supports Deep and Exhaustive effort levels; source and output filtering are not available here. The output this page's human gate reviews before it touches a decision.

Frequently Asked Questions

What can FlowRunner do with You.com?

FlowRunner agents can run Web Search, Get Contents, and Run Research in You.com, plus 3 more actions.

Does connecting You.com to FlowRunner require OAuth?

No. You.com connects to FlowRunner with an API key, no OAuth flow required.

Can You.com trigger a FlowRunner workflow automatically?

You.com doesn't currently expose triggers in FlowRunner. It connects as an action step inside workflows started by another trigger.

Start building with You.com

$100 in credits. No card required. Connect in minutes.

Start Building Free Talk to us