Scribd Classifies 400M Documents With Gemini

Scribd used Gemini 2.5 Flash Lite and Gemini Enterprise batch prediction to classify more than 400 million user-uploaded documents, over 12 billion pages, across Scribd and Slideshare in months. Native PDF input processed over 99% of the corpus without OCR, and batch pricing at 50% below interactive rates made the trust and safety backfill economically viable.