Gamut
VC & InvestingVenture Capital

VC Thesis Screener Agent

Screen every inbound company against your fund's thesis each night and wake up to a sorted fit list.

Updated

Connects with

GmailGmailCrunchbaseCrunchbaseTechCrunchTechCrunchLinkedInLinkedIn
Download template

The problem it solves

Inbound deal flow piles up faster than any analyst can triage it, and most of it never matches the fund's thesis anyway. This agent reads the companies you collect through the day, researches each one, and screens them against your sectors, stage, geography, check size, must-haves, and dealbreakers, so you start each morning with a clear fit list instead of an hour of manual Googling.

Who it's for

  • VC analysts at early-stage funds triaging inbound from founders, scouts, and syndicates
  • Partners who want a daily thesis digest without spending an hour on it
  • Solo GPs and emerging managers without an analyst team
  • Pre-seed and seed funds with a defined, tightly-scoped thesis
  • Investment associates running first-pass diligence on cold inbound

What it does

  1. 1

    Collect companies in your inbox

    Drop company names and one-line descriptions into a plain-text inbox throughout the day, one per line.

  2. 2

    Run on a nightly schedule

    Each evening (default 6 PM weekdays in your timezone), the agent reads every non-comment line from the inbox.

  3. 3

    Research each company

    It looks up each name on Crunchbase, TechCrunch, and LinkedIn to enrich it with stage, funding, location, and notable investors.

  4. 4

    Screen against your thesis

    It checks sector, stage range, geography, check size fit, must-haves, and dealbreakers in order, stopping at the first criterion that fails.

  5. 5

    Assign a verdict and reason

    Each company gets a FIT or NO-FIT call with a one to two sentence reason, and standout positives are noted as signals.

  6. 6

    Flag borderline cases

    Companies that miss a single non-dealbreaker criterion by a small margin are surfaced separately with a note on what would change the verdict.

  7. 7

    Email a sorted report

    It builds an HTML digest with FITs first and the strongest signals at the top, then sends it to your configured address.

  8. 8

    Archive and reset

    Every verdict is saved to a local JSON log for your records and the inbox is cleared so nothing is screened twice.

Key benefits

  • Eliminates the nightly triage of reading, Googling, and cross-referencing every cold inbound
  • Applies your exact thesis criteria consistently, instead of relying on a deck or someone's memory
  • Surfaces the few companies worth a first call ahead of the noise, ranked by signal strength
  • Catches near-misses as borderline rather than silently dropping them
  • Keeps a dated verdict history you can revisit when a passed company resurfaces
  • Runs unattended on a schedule so deal flow never backs up over a busy week

Sample use cases

An analyst pastes 12 founder cold emails and scout referrals into the inbox over the course of the day.

At 6 PM the agent researches all 12, screens them, and emails a digest showing 3 fits with reasons and signals, 8 no-fits with the criterion each missed, and 1 borderline to review.

A solo GP with no analyst team gets a steady trickle of syndicate and LP-referred deals.

The agent handles first-pass deal sourcing automatically each night, so the GP only spends time on companies that already clear the fund's thesis.

A company looks strong but raised at Series A while the fund tops out at Seed.

The agent flags it borderline rather than fit or no-fit, noting that stage is the only miss so the partner can make a judgment call.

A web lookup returns nothing useful for an obscure startup.

The agent screens it on the inbox description alone, marks the verdict as based on limited data, and never delays the run waiting on research.

Key integrations

  • Gmail

    Connected during setup to send the nightly HTML screening report to your inbox.

  • Crunchbase

    Looked up to enrich each company with stage, total funding, and notable investors.

  • TechCrunch

    Searched for recent press and funding announcements about each company.

  • LinkedIn

    Checked for founder backgrounds, headquarters location, and company details.

The screener is built for funds with a defined thesis. The tighter your sectors, stage range, check size band, and short list of must-haves and dealbreakers, the sharper the fit and no-fit calls become. You can re-run onboarding anytime to adjust your thesis, delivery email, or schedule without duplicating your setup.

Because it errs toward NO-FIT when a data point is genuinely ambiguous, the agent favors false negatives you can re-evaluate over false positives that clutter the list. Every verdict carries a concise reason, so reversing a call takes seconds rather than a fresh round of research.

Getting started

  1. Import the workspace — Load the VC Thesis Screener workspace into Claude Code to install the agent and its onboarding skill.
  2. Run agent onboarding — Type /agent-onboarding for a five-minute interview that captures your fund thesis, connects Gmail, and schedules the nightly run.
  3. Add your first companies — Paste a few companies into the inbox file, one per line, and let the next scheduled run produce your first screening report.

Frequently asked questions

What does the VC Thesis Screener agent actually do?

It automates first-pass deal sourcing for venture capital by reading the companies you collect each day, researching them, and screening them against your fund's thesis. Every evening it emails a sorted fit/no-fit report so you skip the manual triage.

Does the agent act on deals without my approval?

No. It only researches, screens, and reports. It never contacts founders or commits the fund to anything. You still decide which fits to pursue, and the report gives enough context to override any verdict quickly.

Which systems does it work with?

It connects to Gmail to deliver the nightly report and uses web research across Crunchbase, TechCrunch, and LinkedIn to enrich each company. Companies are fed in through a simple plain-text inbox you can update anytime.

How is this different from manual screening or generic deal sourcing tools?

Most venture capital deal sourcing software helps you find companies; this agent triages the inbound you already have against your specific thesis. Unlike doing it by hand, it applies your exact sector, stage, check size, and dealbreaker rules consistently every night without you reading and Googling each one.

What happens when it can't find data on a company?

It never blocks the run. If a web lookup returns nothing useful, it screens on the inbox description alone, flags the verdict as based on limited data, and leans NO-FIT on anything it can't confirm so nothing is falsely passed.

Is the agent free to use?

The template itself is free to import and configure. You run it in your own Claude Code workspace and connect your own Gmail account, so there is no separate subscription for the screener.