Vercel Released AI SDK 7 with New Tools for Managing AI Agents

Vercel released the seventh version of the AI SDK — an open-source TypeScript toolkit for developing AI applications and agents, used for over 16 million downloads per week; it also serves as the foundation for Vercel's open-source agent framework, eve. The release adds reasoning management, support for MCP Apps, and extends beyond text agents to voice and video.

What's New for Agent Development

AI SDK 7 standardizes reasoning management: most top models support customizable reasoning, but each API provider implements it differently. Now the reasoning option in generateText and streamText functions is set with a single line of code (e.g., reasoning: 'high') and automatically maps to the provider's native settings — while still allowing manual provider-specific parameters for finer control.

• Over 16 million AI SDK downloads per week
• Single reasoning option instead of different provider-specific settings for each model
• Typed toolsContext — data like API keys are passed directly to a specific tool via contextSchema, never reaching the model's prompt and not leaking to other tools
• Update via pnpm add ai@latest; migration from version 6 via npx @ai-sdk/codemod v7 or skill npx skills add vercel/ai --skill migrate-ai-sdk-v6-to-v7

What's Added for Running and Monitoring Agents

Beyond development, AI SDK 7 extends agent deployment in production: tool call approval, failure resilience via WorkflowAgent, timeouts, and support for sandbox environments. The SDK can integrate with any third-party agent harness — Codex, Claude Code, Deep Agents, OpenCode, or Pi.

For monitoring agent behavior, telemetry, Node.js tracing channels, lifecycle events, and performance statistics have been added. Finally, AI SDK 7 goes beyond text agents: provider-agnostic support for real-time voice agents and video generation has arrived.

What This Means

AI SDK 7 shows where the agent tools industry is heading: from simple prompt-based prototypes to production-grade components with reasoning control, auditable tool access, failure resilience, and comprehensive telemetry — without which an agent cannot be safely exposed to real users.

FAQ

How do I upgrade from AI SDK 6 to AI SDK 7?

You can run automated migration via npx @ai-sdk/codemod v7, which rewrites your code with minimal changes, or use the skill: npx skills add vercel/ai --skill migrate-ai-sdk-v6-to-v7.

What is toolsContext in AI SDK 7?

It's a typed context set separately for each tool via contextSchema — for instance, an API key is passed directly to the tool's exec function, not through the model's prompt, and other tools cannot access this context.

Which agent frameworks does AI SDK 7 support?

The SDK can integrate with any agent harness, including Codex, Claude Code, Deep Agents, OpenCode, and Pi.