Show HN: AI that extracts brand identity from websites to generate ads https://ift.tt/cEsH0iq

Show HN: AI that extracts brand identity from websites to generate ads I built this because I kept procrastinating on creating ads for my projects. The technical challenge was interesting: how do you teach AI to extract "brand identity" from a website? Turns out websites are messy. Finding the actual logo vs random images, identifying brand colors vs generic link colors, understanding brand voice from homepage copy. The solution: Custom vision models + CSS parsing + GPT-4 for voice analysis. You paste a URL, it extracts brand elements, generates platform-specific ads. Not trying to "disrupt advertising" or anything dramatic. Just solving the specific problem of "I need a Facebook ad but Canva makes me want to cry." Built with Next.js, custom image processing pipeline, OpenAI API. The brand extraction accuracy is around 85% for well-structured sites, lower for sites that are... creative with their CSS. Happy to discuss the technical approach or share code snippets if anyone's curious about the brand extraction pipeline. https://board.ad https://www.board.ad June 8, 2025 at 12:15PM

Komentar

Postingan populer dari blog ini

Show HN: Guish – A GUI for constructing and executing Unix pipelines https://ift.tt/HrXz5ub

Twin Peaks for All: Survey Results

Launch HN: Stacker (YC S20) – Create Apps from Airtable or Google Sheets https://ift.tt/3i3ZJso