Methodology

How Forge AI analyzes an idea or a URL.

A single-answer view of the pipeline. No hype — the exact stages, the data sources, and where the model can be wrong.

Idea validation pipeline

  1. Input framing: normalize your one-liner into audience, job-to-be-done, and outcome.
  2. Competitor scan: query public data for adjacent tools; extract positioning and gaps.
  3. ICP synthesis: derive 2–3 concrete personas with pains, triggers, and where they hang out.
  4. MVP scope: smallest testable feature set + explicit out-of-scope list.
  5. Builder prompts: long-form, paste-ready prompts for Lovable / Cursor / v0.
  6. Growth plan: Reddit / LinkedIn / X / SEO strategies with example posts.
  7. Risk & kill criteria: what would falsify this idea in 30 days.

URL analysis pipeline

  1. Server-side fetch: the URL is fetched from our backend, never your browser.
  2. Extraction: meta tags, headings, primary CTAs, above-fold copy, and visible trust signals.
  3. Conversion audit: hero clarity, CTA specificity, friction, and social proof gaps.
  4. SEO diagnostic: title/description length, H1 uniqueness, internal links, schema.
  5. Positioning read: what a first-time visitor thinks you do vs. what you meant.

Honest limits

  • Outputs are reasoned estimates, not guarantees. Validate with real users before building at scale.
  • URL analysis reads server-rendered HTML. Fully client-rendered pages give thinner results.
  • Competitor data reflects what's publicly indexable — we don't scrape gated content.
  • Growth content is a starting point. You still need to press "post".
View a sample reportFAQ