Cloudflare's Clairic pulls data from documents faster and cheaper than LLMs
decoder ·
Cloudflare has introduced Clairic, an AI model designed to extract data from documents more quickly and cost-effectively than traditional large language models. Available in preview on Workers AI, the system allows users to specify desired fields, such as order numbers or invoice totals, and returns the corresponding values with confidence scores in a single pass rather than generating text token by token. The model is intended to function within larger agent chains alongside other System One models like Jev and GLM-5.3, handling tasks such as simple calculations and estimates. In a demonstrated complaint email workflow, Clairic extracted key details while subsequent models applied rules and drafted responses, completing the process in 19.3 seconds compared to 61.7 seconds for GLM-5.3 alone. This approach also resulted in significant cost savings, with the LLM-only method costing nearly twice as much. While Clairic currently struggles with longer chains of calculations, developers are advised to handle those complex computations through their own code. Open weights for the model are expected to be released soon.
decoder