Find the gap.
Build the better page.

Yes. The open-world classifier separates subject from publishing format, so a politics newsroom, sports blog, technology news site or travel video blog can be labeled by both topic and format. If a niche is not in the built-in taxonomy, GapForge creates a dynamic subject from repeated page evidence instead of forcing a wrong category.

No. They are best-fit research candidates based on detected subject/site type and optional live-search results. Verify direct competitive overlap before comparing them.

The percentage uses completed mechanical pass/fail checks. Review items are excluded from the score, so they are shown separately and are not counted as confirmed failures.

No. Public HTML can show quality and coverage signals, but actual traffic causes require rankings, query data, backlinks, brand demand, distribution and analytics evidence.

Validate audience relevance, intent and real demand. If it fits, build original coverage with stronger clarity, examples, trust signals and internal links. Do not copy wording or structure blindly.

Fix confirmed failures on your own site first, manually verify review items, improve important landing pages, research high-fit gaps, build original supporting content, and measure the effect in Search Console and analytics.

Yes as a quality benchmark when the weakness is a confirmed failed check. Review-only signals are investigation leads and should not be treated as proven competitor problems.

Yes. Configure the server-side service account and share the own-site Search Console property with it. GapForge then shows measured 28-day query/page evidence instead of estimates.

Paid logged-in users can save a competitor project. After a Hostinger cron job is configured, GapForge can run daily, weekly or monthly deeper snapshots, keep history and optionally email meaningful changes.

The scheduled deep crawler reads public HTML and can merge samples up to 50 pages per site. Client-rendered content needs the optional render bridge; GapForge keeps raw HTML and rendered evidence separate.

Those are generic one-word observations, not reliable content opportunities. GapForge prefers 2–3 word title/H1/heading phrases and requires repeated competitor-page support before a phrase can become a BUILD candidate.

GapForge does not bypass CAPTCHAs, WAFs or bot challenges. If the SocialKit API itself is intercepted by a verification page, the browser now reports a security-verification error instead of trying to parse the challenge HTML as a comparison result.

STRICT NICHE TRUST MODE blocks the competitor gap report. GapForge shows the detected exact niche/model mismatch and keeps only your own-site evidence plus fetch logs. It does not create competitor weaknesses, topic gaps, differentiators or BUILD tasks from that mismatched pair.

No. It is the deepest interactive subject-focused sample. GapForge only says full small-site coverage is verified when every discovered crawlable URL in that small sample is actually read. Larger sites remain clearly labeled as deep or partial samples.

Yes. A discovery card is a candidate prior, not proof. If fetched pages are ambiguous or too generic, GapForge shows Niche unverified and keeps BUILD paused instead of falsely calling the site a mismatch or a verified competitor.

It uses task-equivalent evidence rather than exact marketing phrases. For home project cost sites, examples include replacement cost, cost to install, cost to repair, installation cost, labor/material cost, project estimates and cost guides. Strong equivalent evidence can verify the same niche.