How it works

From URL to cited memo.

You paste a company’s web address. An AI research agent runs the same research path an analyst would — across the public web only — and returns a structured memo where every claim links back to the page it came from. This page explains what happens in between.

memo depths — Overview (5 credits) and Full (10): 2
sections, in the order an IC reads them: 9
from any claim to the source behind it: 1 click

01 Sources

What the agent reads

Everything comes from the public web — no paid databases, no private data, nothing behind a login. For each company the agent works through the source families an analyst would open in nine browser tabs:

The company’s own site Homepage, product and pricing pages, and the pages its sitemap points to — what they sell, to whom, for how much.
App stores App Store and Play listings: ratings, review velocity, install signals where shown.
Review platforms What customers say on the platforms that matter for the category — praise and complaints alike.
News & press Funding announcements, launches and coverage that signal momentum.
Search results Targeted queries for funding rounds, competitors and published market-size research.
Answer engines How the company surfaces in AI-powered search — an early read on its visibility.

02 Synthesis

How the memo is assembled

The gathered evidence is synthesised into nine sections in the order an investment committee reads them: investment view, business profile, founders & team, market sizing, competitors, funding history, growth momentum, customer sentiment, and risks. Every memo uses the same structure, so a stack of them reads as a comparable cross-section rather than ten differently-shaped documents.

The market-sizing section grounds TAM / SAM / SOM in published research and is explicit about regional splits. The competitor matrix names the top five with positioning notes and the target’s edge over each.

03 Evidence

Every claim carries its source

Each fact in the memo links to the page it came from — a funding figure points at the announcement, a sentiment claim points at the reviews, a market number points at the research it cites. You verify any line in one click instead of taking the synthesis on faith. That’s the difference between a research memo and a summary.

04 Honesty

What happens when data is thin

Early companies often have no disclosed rounds, few reviews, and no published market study with their name in it. The memo says exactly that. It never fabricates a round, a competitor or a metric to fill a section — a gap in the public record shows up as a flagged gap, because “nothing public exists” is itself a diligence finding. Failed runs are never charged.

05 Depth

Overview or Full

5
Overview memo — 5 credits
The triage read: enough structured, cited research to decide whether a company deserves your hour. Built for cold-pitch triage and first passes over a batch.
10
Full memo — 10 credits
Every module, including the deeper research — the version you take into an IC discussion or attach to a syndicate note. Plans start at $129/month; see pricing.

06 Boundaries

What it deliberately doesn’t do

No paid databases.No Crunchbase Pro, no PitchBook seats, no paywalled content — if it isn’t public, it isn’t in the memo.
No private data. Only public web pages; the agent never goes behind logins or paywalls.
No investment advice. The memo is research that supports your judgement — it never tells you to invest.
No invented facts. Thin public record means flagged gaps, not plausible-sounding filler.

07 Read next