Validator
- 8 Devlogs
- 29 Total hours
Startups need to test the market before they can make their dream a reality, Validator is here to help.
Ship #1
Pending review
AI-powered startup idea analyzer. Given a one-line pitch, it returns a structured analysis: competitors, market size, weaknesses, pricing suggestions, and a computed likelihood-of-success score.
Features
Likelihood of success, computed deterministically from sub-scores (not just an LLM guess)
Monetization score and technical difficulty rating
Market size and growth rate, verified against real published figures where possible
Competitor list with web-scraped details (website, description, pricing hints)
Weaknesses and risk severity
Pricing model suggestions with rationale
Tech Stack
Backend: Flask (Python)
AI: Groq API (Llama 3.3 70B)
Market data: scraped search results (Mojeek), parsed for dollar figures and CAGR
Competitor scraping: BeautifulSoup, Requests
Frontend: HTML, CSS, JavaScript
18m logged
Shipping now!
15m 25s logged
UI issue on the vercel, fix then push.
7h 16m 42s logged
I added sources so you can tell where it got the data from, the issue was that edge and chrome were rejecting requests so I had to go a seperate route
Found an issue with the model will all of the characteristics clustered around 50… Will need to fixthat.
5h 40m 32s logged
Implemented other insights and increased the different types of monetization models the model is trained on so it can make the best decision.
3h 50m 34s logged
Working on implementing the mathematics into my model. So far so good, I will now look at improving other insights and adding sources…
4h 30m logged
Did some math, to calculate the success index. I realized that the LLM returns the ingredients (market size, growth, monetization, difficulty, competitor count, weaknesses).
So I have to normalize the signals from 0-1,
market = clamp( (log10(TAM_dollars) - 7) / 4 , 0, 1 )
growth = clamp( CAGR% / 30 , 0, 1 )
n = number of competitors
n = 0 → 0.40 (unvalidated market)
n = 1–3 → 0.90 (validated, room to enter)
n = 4–6 → 0.70
n = 7–9 → 0.50
n ≥ 10 → 0.30 (saturated)
feasible = (10 - difficulty) / 9
success% = round( 100 / (1 + e^(-6 * (S - 0.5))) )
Now working on implementation
6h 29m 59s logged
Working on this validator, so far I have used knowledge from previous projects to get the website scraping to work, I am just not trying to get the predictions to work… Very rough right now, it may take a while for that to start working.
36m 26s logged
Starting to devlop this startup validator to give people a glimpse into how well their ideas will do on the market. Inspired by the bad sharktank pitches.
