@@ -59,8 +59,24 @@ <h1>Clone a development database in seconds, not minutes.</h1>
5959 </ div >
6060 </ div >
6161
62- < div class ="hero-visual ">
63- < img src ="./assets/dbfork-demo.gif " alt ="dbfork demo showing local database branching " />
62+ < div class ="hero-proof ">
63+ < div class ="compare-card strong ">
64+ < h3 > dbfork</ h3 >
65+ < p class ="metric "> 1 to 3 seconds</ p >
66+ < p > Server-side template clone, built for repeated use during active development.</ p >
67+ </ div >
68+ < div class ="compare-card ">
69+ < h3 > pg_dump | pg_restore</ h3 >
70+ < p class ="metric "> 2 to 5 minutes</ p >
71+ < p > Good for backup-style workflows, too slow for every migration and feature branch.</ p >
72+ </ div >
73+ < pre class ="hero-terminal "> < code > $ dbfork create feature_migration
74+ created feature_migration from postgres_dev
75+
76+ $ dbfork diff feature_migration
77+ tables changed: 3
78+ columns added: 5
79+ indexes changed: 1</ code > </ pre >
6480 </ div >
6581 </ section >
6682
@@ -119,17 +135,21 @@ <h2>Use the local database like code, not like a fragile shared pet.</h2>
119135 </ p >
120136 </ div >
121137
122- < div class ="compare ">
123- < div class ="compare-card strong ">
124- < h3 > dbfork</ h3 >
125- < p class ="metric "> 1 to 3 seconds</ p >
126- < p > Server-side template clone, built for repeated use during active development.</ p >
127- </ div >
128- < div class ="compare-card ">
129- < h3 > pg_dump | pg_restore</ h3 >
130- < p class ="metric "> 2 to 5 minutes</ p >
131- < p > Good for backup-style workflows, too slow for every migration and feature branch.</ p >
132- </ div >
138+ < div class ="workflow-grid ">
139+ < article class ="feature ">
140+ < h3 > Branch fast enough to use it daily</ h3 >
141+ < p >
142+ The performance win matters because developers only keep a branch workflow
143+ when it feels as disposable as a git branch.
144+ </ p >
145+ </ article >
146+ < article class ="feature ">
147+ < h3 > Inspect before merge</ h3 >
148+ < p >
149+ Compare the branch against the source database, review drift, and only
150+ keep the experiment when the schema change is real.
151+ </ p >
152+ </ article >
133153 </ div >
134154 </ section >
135155
0 commit comments