Hey Product Hunt! 👋

With Claude Code and Codex, shipping an MVP takes a weekend. The real trap now is spending two weeks building something, only to realize 20 funded competitors already own the SEO keywords—or nobody wants to pay for it. And asking a standard LLM "is my idea good?" usually just gets polite flattery.

I built EarlyLore so builders can check the real market before writing code:

1. Competitor-grounded idea roast: pulls 20 live competitors' monthly web traffic and entry pricing first, then runs an 8-seat roundtable of skeptical buyers who challenge your assumptions.

- Sample report ("AI SMS receptionist for dental clinics", no login needed): https://earlylore.com/evaluations/24510e6c-7aa4-4f53-b5fa-b6028b581331

2. 445,944 software products across 249 niches: includes a "Rising" filter (4,442 products) ranked by 12-month traffic growth slope rather than total legacy traffic.

3. 40,900+ homepage CSS profiles + MCP: extracted real computed DOM font stacks, type scales, and 5-color ratios from 40k+ live landing pages so coding agents have real design references instead of defaulting to purple gradients (https://earlylore.com/design).

One fun finding from cross-tabulating the dataset: among products with verified revenue disclosures, 26% actually have under 10,000 monthly visits, and 67% get their primary traffic from organic search (SEO).

Would love your honest feedback!