Book a demo

Trawl

Send Vernais to get the facts. It brings them home.

Research the open web in plain language.

The old way

Research used to mean tabs, contractors, and copy-paste.

Market and competitor research lives outside your tools. You open forty tabs. You skim reviews. You hand a brief to a contractor and wait a week. You paste the findings into a doc that nobody connects back to the decision.

The facts you gathered never touch the data you already have. Vernais closes that gap: it lands external research in the same brain that reasons over Stripe, Jira, and Zendesk. External facts and internal facts, side by side, in one investigation.

01 · What it is

A conversation, not a config screen.

You don’t fill out a form. You have a conversation. Ask for what you want; Vernais reverse-prompts for the settings that matter, gathers the URLs, shows you what it found, and waits for your go.

  1. 01

    You ask. “Scrape reviews of Northwind.” That’s the whole request.

  2. 02

    Vernais asks back. How many pages per source? How deep: light, medium, or thorough? It surfaces the choices as tappable cards, not a settings panel.

  3. 03

    Vernais gathers the URLs. It runs quality search across multiple engines and assembles a real source list.

  4. 04

    Vernais splits public from paid. Free-to-read sites go in one bucket; paid or API-gated sources it names explicitly, so you decide. By default, Vernais skips the paid ones.

  5. 05

    You confirm. Only when you tap go does a single byte get scraped.

  6. 06

    It runs, and you watch. A live status widget streams the count.

02 · Quality first

Real sources, not word-match junk.

A scrape is only as good as the pages it reads.

Most scrapers grab whatever a single search engine word-matches. Ask for “best running shoes” and you get Best Buy, a dictionary definition of “best,” and a listicle of restaurants. Vernais gathers SEO-grade URLs the way a careful researcher would.

Multiple engines, merged

It queries several search engines at once and interleaves the results, so the strongest sources rise regardless of which engine found them.

Junk filtered out

Dictionary pages, definition sites, and known-noise domains are dropped before anything is read.

Diverse by domain

No single domain can dominate the list, so one official page can’t drown out the community and review sources that count.

Real sources

Reddit, Trustpilot, G2, official sites, forums, and the outlets people actually cite.

03 · Live status

Live and controllable.

You’re never watching a black box. The moment a scrape starts, a status widget appears in the chat and updates as it works.

  • Running or done. The widget shows the current state at a glance.
  • Items, pages, sources. Live counts as the crawl progresses.
  • Recent activity. The latest pages read, streaming in.
  • Stop any time. Say “stop” and it stops. A stop is a stop, never mistaken for a new search.

A scrape is thorough, not instant. Many sources across many pages take real minutes. An early status showing a handful of items is progress, not a stall.

04 · The Company Brain

It joins your brain.

This is the part standalone scrapers can’t do. Scraped results don’t sit in a separate export. They become searchable evidence inside Vernais, right next to your connected data.

  • Answers on what you found. “From the scraped reviews, what are people complaining about?” Vernais answers over the corpus you just gathered.
  • One reasoning surface. External facts and your own Stripe, Amplitude, and Zendesk data are reasoned over together: the same chat, the same lane discipline.
  • Clear separation. The AI always knows which facts came from the open web and which came from your connected tools. External is external. Internal is internal. It never blurs the two.

05 · On demand

Your research analyst, on demand.

Think of it as a researcher who never gets tired of tabs.

01

You brief it in a sentence

No query syntax, no scraper config.

02

It proposes the plan

Sources, depth, public vs paid, surfaced for your approval.

03

It does the reading

Dozens of pages across dozens of domains, so you don’t.

04

It reports back where you work

Findings land in chat, ready to question, and ready to become a Signal, a Hypothesis, or an Initiative.

The web stops being a place you visit. It becomes a place Vernais brings to you.

The whole picture

How the pieces fit.

The Trawl lane sits beside Data Lookup, Deep Investigation, and Web Search: one surface, the right tool per question.

01

Conversational, confirm-first

You ask, Vernais reverse-prompts, you approve, then it runs.

02

Quality URL gathering

Multi-engine, junk-filtered, domain-diverse, SEO-grade sources.

03

Live status widgets

Running / items / pages / sources, stoppable any time.

04

Joined to the Company Brain

A searchable corpus you can question next to your own data, external clearly marked.

05

Part of the same chat

One surface for every question, the right lane picked per question.

The web, brought to you

Ask for the research; the open web lands where you already work.