Send a prompt to a hosted AI API and it arrives on someone else's machine as readable text. Staff. Logs. The routing layer. The company that owns the model. Nothing in the architecture stops any of them from reading it. What stops them is a policy. Every model served on…

The TEE stack and the encryption layer are open source, so the claim is checkable: github.com/chutesai/sek8s Link GitHub - chutesai/sek8s: Secure, standalone k8s setup with TEE. Secure, standalone k8s setup with TEE. Contribute to chutesai/sek8s development by creating an…

New on Chutes: docuextract. Feed it a PDF or an image. It hands back clean markdown or schema-guided JSON. Invoices, contracts, bank statements, scanned forms, research papers. Multi-page documents in a single call. And like 100 percent of the models served on Chutes, it…

Try it in the playground: chutes.ai/app/chute/vonkaise… Under the hood, three OCR engines sit behind one endpoint. Between them they cover fast page-to-markdown OCR, schema-guided JSON extraction, template generation, and one-shot multi-page parsing. Link docuextract -…

Week 13. The rundown: → Chutes is migrating from the current smart contract mechanism to conviction lock, enforced at the protocol level. More announcements next week → Cached prompt tokens: 90% off, up from 50%. Long-context and repeated calls just got a lot cheaper →…