Sable is private, verifiable compute for AI agents. Run models, deploy agents, use tools and execute workloads with privacy, spending controls and cryptographic proof built into the stack.