LiberPixels
Smart Free Image Search For Creatives
Abstract
Flow Communications, a South African integrated communications firm, needs its designers, developers and content creators to find licence-compliant free images quickly. Platforms such as Unsplash, Pexels and Pixabay work in isolation, give no help to users who struggle with keywords, and show licence terms inconsistently, which wastes time and creates copyright risk. The solution also had to be built using free tiers only.
LiberPixels is a browser-based platform that lets users describe what they need in plain language and returns results from seven providers (Unsplash, Pexels, Pixabay, Openverse, Freesound, Wikimedia Commons and Iconify) in one grid, with licence information on every result. It uses a three-layer architecture: an HTML, CSS and vanilla JavaScript front end, Node.js serverless functions hosted on Vercel, and a data layer combining Firebase (authentication and Firestore) with a MongoDB Atlas cache that cuts repeat API calls. Gemini powers image-based search and keyword extraction, while Groq drives Pixie, a conversational assistant that asks follow-up questions to refine vague queries. Gateway packages isolate each external provider so it can be replaced independently.
Key features include multi-source aggregation, filtering and sorting, category browsing, AI-assisted and reverse image search, user accounts, shareable collections with role-based access, and a built-in image editor. Input and output validation protects data integrity, and risks such as quota exhaustion, provider downtime and API key leakage are addressed through caching, fallback providers and server-side secret storage. LiberPixels shows that a unified, guided search experience can reduce the effort of sourcing compliant assets while running at no cost on free infrastructure.
Videos 1
Watch presentations, demos, and related content
LiberPixels System Demo
Like, comment, and subscribe on YouTube to support the creator!
LiberPixels System Demo
Gallery 1
Explore the visual story of this exhibit
Title Image