Product teardown
By coExploro · Captured · Updated
Reviewed and edited by coExploro · Independent, not affiliated · Public methodology · Corrections: support@coexploro.com
How this page was made: Drafted by four AI analysts from a capture of supabase.com on . Checked by coExploro’s founder on .
A curated sample showing what coExploro observes about public homepages, pricing pages, feature pages, and UX signals.
Illustrative public sample. coExploro is not affiliated with Supabase.
Sample output
Overall score 90/100
PM 91
UX 88
Summary Card + findings
Summary Card
This public version uses the same compact card pattern as private teardowns: one score, four persona signals, and the strongest findings compressed into a card a product team can forward.
Read the methodologyOverall score
90/100
Confidence
high
PM
91/100
UX
88/100
teardown · captured 2026-08-10
Top findings
Findings
Marketing
High impact
A public repository star count in the primary navigation swaps conventional logo-cloud credibility for a signal the target audience knows how to interpret without explanation.
PM
High impact
The footer carries "Switch from Firebase" and "Switch from Neon" as their own links, alongside audience paths for AI builders, agencies and startups. Rival product names are treated as navigation, not only as comparison content.
UX
High impact
Free at $0, Pro from $25 and Team from $599 a month, plus Enterprise on custom pricing, are compared across about sixty feature rows in eight groups. Every capability is listed rather than summarised.
Design
Medium impact
An interactive usage calculator directly under the tier grid doubles the decisions on one screen: clarity for a curious reader, load for a scanning one.
FAQ
This sample covers pages that were publicly visible on . No account, private data or share link is involved. It carries observations only. The underlying run does produce recommendations, but they are written for someone deciding how to compete with the product, which is not what a public page about that product should say.
All six UX dimensions were observable in this capture. Where a loading state never appears in frame, perceived performance goes unscored and confidence drops a band. It is a property of the run, not a verdict on the product.
That all six UX dimensions were observable in this capture. Most runs land a band lower because perceived performance needs a loading state to appear in frame. It describes the evidence available, not the quality of the product.
Next