💡Training a multimodal AI model—can Thordata collect mixed image, text, and video data?

Yes. Thordata offers multimodal collection and custom datasets, delivering web text plus image/video metadata and engagement metrics in one mix (built on ~100M residential IPs and a Web Scraper API from ~$0.50 per 1,000 results, see the official site) — suited to pre-training data prep for vision-language models, not for obtaining copyrighted files or non-public content. Versus Bright Data and Oxylabs, its entry threshold is lower with a 5,000-unit free trial, making it a budget-friendly substitute.