AI SDK is the TypeScript SDK for building AI applications, features, frameworks, and agents across any model provider. AI SDK 7 focuses on what it takes to run AI in production.

Vercel is a platform for deploying modern web applications with speed and simplicity, offering developers a seamless experience for hosting and scaling their projects. With features like instant deployment, serverless functions, and global CDN, Vercel empowers developers to focus on building great user experiences without worrying about infrastructure management.

Vercel

Vercel releases AI SDK 7, a major update to its TypeScript SDK for building AI applications and agents. Key additions include: standardized reasoning control across providers, typed tool and runtime context, provider file and skill uploads to avoid re-sending large inputs, MCP Apps support with sandboxed iframes, a terminal UI for quick agent testing, durable/resumable WorkflowAgent, granular timeout configuration, SandboxSession abstraction, HarnessAgent for integrating Claude Code/Codex/Pi and other harnesses, revamped telemetry with single-registration OpenTelemetry support and Node.js tracing channels, per-step performance statistics, lifecycle callbacks, provider-agnostic realtime WebSocket support, and experimental video generation. Migration from v6 is automated via a codemod.

AI SDK 7 is now available

Nice update. The provider-agnostic approach is what I’m most interested in—it makes switching models much less painful.

At this point Vercel is more of an AI company than a web company