Abdi Bedel logo
Available · Barcelona
← All productsFROM THE LABWK 01 OF 26
NEXT.JSSUPABASESTRIPEAIFOUNDER & PRODUCT LEAD

SignalCheck

The AI that challenges your idea before you build the wrong thing.

SignalCheck is a reality-check tool for founders. You paste an idea. It probes your thinking, researches competitors, and hands back an Honest Brief with the questions you haven't answered and what to validate in seven days.

Try SignalCheck ↗Read the story ↓

— STATUS

Live

— DAY-1 USERS

0

— FIRST STRIPE HIT

Pre-revenue

— NEXT

Week 2 measurement

— By the numbers

Web · live data

— Core

— Launched

2026-04-18

— Signups

0

— Retention

—

— MRR

£0

— Stack

Next.js · Tailwind · Supabase

— Primary metrics

— Visit → Signup

—

— Daily Active Users

—

— Activation Rate

—

— Share Rate

—

— Secondary · Web

— Median Session

—

— Lighthouse

—

— Bounce Rate

—

— Organic %

—

— The story

I built SignalCheck because I kept wishing it existed.

I have a paste-box full of product ideas - most of them sound good at 11pm. Half of them have been built already. The rest need a conversation before they need a line of code. The people around me either tell me it sounds good, or ChatGPT tells me it's a great concept. Neither is discovery.

So I built the tool I'd have paid for. And I made it GBP5 because that's what I wished a real discovery session cost when I was trying to figure out what to build next.

It's Week 1 of 26 products in 26 weeks. I used SignalCheck to pressure-test SignalCheck before I built it. That felt right.

— The four decisions that matter

where it would have died if I'd got these wrong

  • No instant output. The user sees the research running. Builds trust. Separates SignalCheck from every generic AI chat.

— The stack

click any tool to see why I picked it

— Frontend3

— Backend1

— AI & Payments3

Next.jsFrontend

App router + server components meant I could ship the visible research step without a separate backend. Server-side streaming for the brief output — felt like magic, was just the framework doing its job.

Shipped Wk 1—

— Week 01 numbers

honest start, no spin

— STATUS
Live
— DAY-1 USERS
0
— FIRST STRIPE HIT
Pre-revenue
— NEXT
Week 2 measurement
— What shipped
Full flow from paste to Honest Brief, Stripe credits, OAuth sign-in
— What changed
Built and shipped in a 12-hour Saturday sprint
— The numbers
Posting end of week

— Week 01 retro

no highlight reel

  • — Worked

    The visible research step is the product. Users watch the research happen and trust the brief more because of it. It's the one decision that makes SignalCheck feel different from pasting the same idea into ChatGPT.

  • — Didn’t

    The probe questions are doing the job, but the jump from probe answers to the final brief needs more work. Some briefs read sharp, some read generic - and the gap correlates with how specific the user's answers were, not the idea itself.

  • — Surprised

    Ship the share card on day one, not day three. The Honest Brief is made to be shared - founders post uncomfortable verdicts - and I underestimated how much organic distribution that one feature would do.

  • — Next

    Wins and failures both. No highlight reel.

Want to try SignalCheck? Live at signal-check-eta.vercel.app.

signal-check-eta.vercel.app → →
← All productsNext: Miguelito →