Storyblocks debuts AI search to fix stock media discovery

If you have ever lost an afternoon digging through page 30 of a stock video library for a clip that does not look like it was filmed in 2012, Storyblocks feels your pain.

The subscription stock platform rolled out an AI-powered search overhaul designed to fix the worst part of video editing: asset discovery.

Pruning the junk and boosting fresh uploads

Instead of just hoarding as many clips as possible, Storyblocks made an unusual move. Even after growing its catalog by 40% in 2023, the platform actively deleted nearly 200,000 underperforming files to clean up search clutter.

To keep results fresh, Storyblocks built a proprietary algorithm that recalculates the performance of newly uploaded assets every four hours. Early testing shows that pushing recent clips up the ranks boosted downloads by up to 60% compared to static rankings.

Under the hood, an in-house model based on Google's BERT framework evaluates user interactions in near real-time. The company also enforced diversity benchmarks, ensuring about 24% of first-page video results feature BIPOC talent.

AI covers, smarter queries, and audio spectrograms

Storyblocks also tapped OpenAI's large language models to interpret fuzzy or flawed search queries. When a creator types a vague prompt, the system now infers creative intent rather than serving a blank page.

Computer vision handles the visuals. The platform uses AI to pick the most engaging thumbnail frame for each clip, which A/B tests proved drove higher click-through rates. It can also identify recurring actors and suggest alternative angles.

For audio, Storyblocks turns songs into visual spectrograms. Its computer vision models then analyze frequency patterns to surface tracks that match the exact vibe of an existing piece.

Why it matters

For video editors and lean marketing teams, this is an immediate quality-of-life upgrade. Hunting down clean B-roll and matching background music drains hours, and smarter discovery pays off right away without altering existing workflows.

We think Storyblocks took a remarkably practical path here. While competitors are chasing generative video moonshots, commercial creators still just need vetted, high-res stock footage without the digging.

The trade-off is aggressive curation. Purging 200,000 files and prioritizing assets that click every four hours makes search punchy, but it risks burying obscure, niche clips that certain creators might eventually need.

Still, doubling down on search utility looks like the right defense against AI video tools that are not quite ready for high-stakes commercial production.