Predicate Ventures
All projects

Liveproduct

Tough Crowd

Feedback from 20 users, before you have 20 users.

Drop in a URL. Synthetic personas research your market, then use your product in live browser sessions, click by click. Each run returns a scorecard, friction findings pinned to the step that produced them, and a ranked fix-list.

  • Synthetic usage, not surveys
  • Real browser sessions
  • Screenshot-pinned findings
  • First run free
Run your first gig free (opens in a new tab)
  • Agentic orchestration
  • NLP & information extraction
  • Data science & evaluation
  • Cloud infrastructure
test files, approximately
101
personas in your free run
3

The problem

Pre-product-market-fit research is too slow and too expensive. So founders ship on instinct, and find out why users churned after they churned.

The approach

Tough Crowd simulates usage, not interviews. The personas skip the questionnaire, bounce off your signup form and report where it happened. The sampling frame mandates abandoners, skeptics and critics alongside adopters, so the happy path cannot dominate by construction.

How Tough Crowd fits together

  1. Your URL

  2. Personas

    Adopters, abandoners, skeptics and critics

  3. Live sessions

    Claude vision driving the real product

  4. Findings

    Pinned to the step, with screenshots

  5. Verdict

    A scorecard and a ranked fix-list

Re-run after a fix and the change shows as a scorecard delta

Synthetic usage of the real product, not synthetic interviews.

What Tough Crowd does

  • Personas bounce off your form

    Instead of answering questionnaires, they use your signup form and report exactly where they got stuck.

  • Adopters, abandoners, skeptics

    The sampling frame mandates critics alongside fans, so the happy path can't dominate by construction.

  • Re-run diffing

    Ship a fix and the next run shows it as a scorecard delta, not a fresh report to reread.

  • Honest about being synthetic

    Synthetic data is a hypothesis, not a finding. The notice that says so is non-suppressible by design.

What’s inside Tough Crowd

  • A worker pipeline: personas, snapshot, live browser session, diffing, evaluation, findings, synthesis, artifacts
  • Claude-vision Chromium sessions driving the real product rather than a recording of it
  • Findings pinned to the step that produced them, with screenshots
  • Re-run diffing, so a shipped fix shows up as a scorecard delta
  • Built on voice-of-agents, pinned by commit and never forked
Synthetic data is a hypothesis, not a finding. The notice is non-suppressible by design.
Tough Crowd, on its own limits

Where Tough Crowd stands today

  • Roughly 101 test files, with CI, deploy workflows and infrastructure as code

The Tough Crowd tech stack

  • FastAPI
  • AWS Lambda
  • ECS
  • React 18
  • Vite
  • TypeScript
  • Tailwind
  • Supabase Postgres
  • row-level security
  • Stripe
  • PostHog
  • Terraform

Try Tough Crowd

First run is free: three personas, one live session. Paid runs are $8 to $50, by crowd size.

Run your first gig free (opens in a new tab)