Preuve.ai Review: differentiated tool, editorial gaps (80/100)

Field	Value
Domain	preuve.ai
Category	Startup validation tools
Pricing	Freemium; $19–$19 USD price range; $29 one-time Founder Report
Pages crawled	40
Crawl date	2026-09-24

Preuve.ai scores 80/100 as an AI-powered startup idea validation tool for founders, serial founders, and solo builders. Its strongest dimensions are positioning at 91/100, audience and messaging at 89/100, performance at 85/100, usability at 85/100, design execution at 85/100, and technical SEO at 82/100.

The product’s central promise is specific: it scans 60+ live data sources, scores an idea, maps competitors, and sizes the market. The main weaknesses are editorial and self-serve depth. Writing quality scores 68/100, editorial QA 64/100, and documentation and self-serve help 45/100. Accessibility scores 78/100, with keyboard-navigation and contrast issues still to address. Performance is strong overall, though the homepage records a 4.1-second mobile LCP and 150ms INP. Review-content integrity is 80/100, with methodology and disclosure gaps noted in the score table.

Preuve.ai makes a clear, differentiated promise: “Get Proof Your Idea Can Sell.” The homepage supports it with a commitment to 60+ live data sources, customer counts, named logos, review scores, and source-linked claims. It also draws a clear contrast with opinion-based validation and alternatives such as ChatGPT and IdeaProof. That combination makes the proposition specific, falsifiable, and easy to distinguish. The remaining friction is comprehension: terms such as “live data,” “viability score,” and “TAM/SAM/SOM” add jargon for non-expert readers. Add a plain-language category label and a short “for founders” clause to the hero while keeping the evidence-led promise intact.

Preuve.ai speaks credibly to founders, with the FAQ naming first-time and serial founders, vibe coders, and solo builders. The site also answers practical questions about cost, trust, and how to begin through pricing, testimonials, and a clear call to action. The gap is placement: the hero says “Get Proof Your Idea Can Sell” without naming the audience or category in plain words. A visitor who does not already know the startup-validation category must infer what the product is. Add “for founders” to the hero and state “Startup Validation Tool” above the fold. Those small changes would make the existing audience definition work earlier in the decision journey.

The homepage guides visitors toward action with a prominent “Try free →” CTA, and core tasks such as finding pricing and starting a trial are straightforward. The pricing page is less decisive. Multiple plans use similar styling, so the “Best deal” offer—10 ideas at 45% off—does not stand out until the visitor reads closely. The $29 Founder Report is visually prominent but is not clearly labeled as a one-time purchase. Mobile navigation adds friction: its hamburger menu contains 34 links without grouping or prioritization. Differentiate the best offer, label the report as one-time, and group mobile links around the main user tasks.

Preuve’s accessibility score is 78/100: the main experience is usable, but keyboard and low-vision users still encounter avoidable friction. The homepage has no skip link to bypass navigation, and Lighthouse reports insufficient contrast for the primary CTA and form inputs. Key interactive elements also lack reliable focus styling. These are contained issues rather than a fundamental barrier, but they affect the first moments of use and make keyboard navigation less clear. Add a focusable skip link targeting main content, restore visible focus states, and validate interactive text at 4.5:1 contrast for normal text or 3:1 for large text. Add accessible labels to icon-only controls as part of the same pass.

Preuve.ai presents a clean, modern interface with consistent spacing, typography, and responsive behavior. The crawl reports no horizontal scroll or tap-target failures on mobile, which supports the strong overall design score. The repeated weakness is light-gray text on white: body copy, pricing text, and the “Entrepreneurs trust us” section use #9CA3AF at 2.8:1, below the 4.5:1 normal-text requirement. Pricing cards also vary between 20px and 24px of padding. Darken secondary text to the tested #4B5563 value, then standardize card padding at 24px. These are focused token and component changes, not a redesign.

Preuve’s performance score is 85/100, with INP at 150ms. The LCP element is the main heading, so the remaining delay is tied to font delivery rather than an oversized above-the-fold image. The heading uses font-display: swap, but the font origin is not preconnected and the critical font is not preloaded. The hero background is also served without explicit dimensions and is not optimized for the mobile viewport. Preconnect and preload the critical font, give the hero container explicit dimensions, and prepare a mobile-sized image variant. End with that concrete sequence: preconnect, preload, dimension, and mobile-size the hero asset.

Preuve’s strongest copy turns an abstract promise into concrete evidence: “Evidence before you build,” live data from 60+ sources, a $29 one-time offer, about ten minutes, and 15 mapped competitors. The editorial risk is proof density without enough qualification. The homepage cites CB Insights’ 42% figure and claims 8,000+ ideas analyzed since 2026, but the source, date, population, and scope should sit beside each statistic. The crawl also reports 1,854 words in one paragraph on the homepage and 4,525 words in one paragraph on /blog, which weakens scanning and extraction. Add claim-level qualifiers, restore semantic paragraph structure, and complete alt text for 20 homepage images and 4 /about images.

Preuve’s comparison pages provide useful decision support through source linking, structured output, and explicit comparisons with competitors. The weak point is the recommendation layer. The “Best Idea Validation Tools (2026)” page places Preuve AI at #1 with a 10/10 verdict, but does not explain the win separately for different audience types; the reasoning is implied through the 60+ live-source claim. The IdeaProof and ChatGPT pages also use more than 10 axes, with some overlap. Add segmented recommendations that state who should choose which option, why, and what tradeoff follows. Then trim redundant axes so source verification, market sizing, and pivot recommendations remain prominent.

Preuve’s comparison content is well sourced in its core claims, but the trust layer is incomplete. The IdeaProof and ChatGPT pages show affiliate links without a visible disclosure near the first link, and neither page provides a dedicated methodology block before its verdict. The criteria are implied in the comparison grid rather than stated plainly. The comparison pages also have Review/AggregateRating schema without visible hands-on testing claims, making the structured presentation overstate the review basis. Add a short method section naming the criteria and evidence basis, and say when no hands-on testing is claimed. Place a plain-language affiliate disclosure at or above the first affiliate link.

Preuve.ai has no detected active traffic suppression, but its March 2026 registration means historical stability cannot yet be assessed. The material structural risk is discoverability: the sitemap contains 247 URLs including numerous .md files, and deep pages serve trailing-slash variants with identical content without redirects. Historical evidence is also limited to 4 Wayback snapshots in 2026. The site operates in a validation and SaaS comparison space exposed to AI Overviews and aggregator displacement, although actual traffic trends were not available. Keep the stability assessment focused on the young domain and limited history; address URL normalization and .md canonicalization in Technical SEO.

The copy is strongest when it follows the product’s own standard: specific claims supported by sources, dates, prices, and concrete outputs. The QA gate remains incomplete around load-bearing statistics. The homepage states that 42% of startups fail because there is no market need and cites a July 2026 benchmark drawn from 6,000+ ideas, but each statistic needs a named source, population, definition, and methodology note. Mechanical QA also reports missing alt text on 20 homepage images, 4 about images, and 19 mcp images. Finally, the markdown article mirror lacks a title, meta description, and H1. Annotate claims, repair the image metadata, and redirect or fully metadata-enable the mirror.

Preuve’s public support surface is a lightweight /support portal linking to the external Intercom help center at https://intercom.help/preuve-ai/en/. The captured portal has no search box or sidebar navigation, and the external center was not crawled, so its article quality cannot be credited here. The primary domain also lacks a native getting-started guide and task how-tos. One positive is AI readability: llms.txt was updated on 2026-09-20 and homepage FAQPage schema covers pre-sales questions about pricing, privacy, and model architecture. Add a native three-step quickstart, topic-based links, and a prominent search path; keep the external center clearly connected so users do not lose context during evaluation.

Preuve.ai has a strong technical foundation: raw and rendered word counts show 0% JavaScript dependency across sampled templates, and HSTS preload, CSP, and nosniff are present. Robots.txt is correctly configured, but the sitemap includes numerous .md files that need canonicalization or exclusion. The main debt is URL normalization. https://preuve.ai/idea-validation/ returns HTTP 200 with identical content to the canonical URL and no redirect chain, while multiple sitemap-listed .md pages return HTTP 200 without canonical tags. /free-tools also exhibits horizontal overflow at a 390px viewport. Enforce one trailing-slash convention with a single-hop 301, then canonicalize or exclude source-file URLs and constrain the mobile layout. Homepage FAQPage schema is present, while the IdeaProof and ChatGPT pages have Review/AggregateRating schema overreach. Field Core Web Vitals remain unavailable in the provided bundle, although the supplied performance audit reports mobile LCP and INP measurements.

Preuve.ai earns its 80/100 through a differentiated, evidence-led proposition and strong execution across positioning, messaging, usability, design, performance, and technical SEO. It is a good fit for founders who want proof before building and need structured startup validation.

Resolve the remaining accessibility issues around keyboard navigation and contrast. Add URL normalization and remove or canonicalize unintended .md sitemap URLs as near-term technical priorities, since they create material indexing risks. The product’s promise is already specific; the supporting evidence and guidance should match it.

This 13-dimension review uses the public site crawl, site facts, and the supplied dimension score table for preuve.ai. The crawl covered 40 pages on 2026-09-24. Scores and summary metrics are preserved from the supplied inputs; this frame does not add findings beyond them.

Google Search Console was not connected, so no Search Console enrichment is included. The score table identifies the primary evidence gaps as claim-level methodology, review-content disclosures, and limited self-serve documentation. See How SiteList scores for the review method and dimension definitions.