Vercel announced several updates to its AI development tools during the Ship AI event. The event included keynotes and talks on AI workflows, agents, and observability.
Among the releases was the beta version of AI SDK 6, which adds an agent abstraction layer for defining and reusing AI agents in projects. This layer allows developers to specify agent behaviors once and apply them across different parts of an application. The SDK also incorporates tool execution approval, integrating human-in-the-loop processes to review and confirm AI actions before they proceed. Type safety extends across supported AI models and user interfaces, ensuring data consistency and reducing runtime errors through compile-time checks.
The Vercel Marketplace received updates to facilitate discovery and integration of AI agents and services. It now supports installation of agents including CodeRabbit for code assistance, Corridor for workflow automation, and Sourcery for refactoring tasks. AI services available through the marketplace cover Autonoma for data handling, Braintrust for testing model outputs, Browser Use for web interactions, Chatbase for chat data analysis, Descope for identity management, Kernel for processing tasks, Kubiks for pipeline operations, and Mixedbread for coordinating multiple models. These integrations use unified billing and simplified setup to connect directly within Vercel projects.
Vercel introduced use workflow, an open-source library for TypeScript. This tool converts standard functions into durable workflows by managing retries for failed operations, executing background steps without blocking the main thread, and providing built-in observability for tracking execution. It operates independently of specific frameworks, allowing use in environments like React, Next.js, or plain Node.js applications. The library handles state persistence and resumption, making it suitable for long-running processes in AI-driven systems.
Vercel Agent, now in beta, serves as an intelligence component for deployed applications. It performs AI-based code reviews on pull requests, generating patches that undergo validation in real-world scenarios before application. The agent also monitors for anomalies in production, such as unexpected performance drops, and initiates automated investigations to identify root causes and suggest fixes. Access to the beta includes promotional credits for usage.
The Vercel Python SDK entered beta, enabling deployment of Python-based web frameworks on Vercel AI Cloud with no manual configuration. It supports FastAPI for API development and Flask for lightweight servers, handling scaling and routing automatically in a serverless setup. Developers install it via pip and deploy projects as they would with JavaScript equivalents.
To support team adoption, Vercel launched An Agent on Every Desk, a program offering guidance on implementing AI agents. It includes consultations to identify suitable use cases, access to reference templates, and assistance in moving prototypes to production environments. For go-to-market functions, an open-source agent template processes leads by enriching data from external sources and qualifying prospects based on criteria like company size or engagement signals. Another open-source template connects Slack channels to SQL databases, enabling natural language queries for business metrics and generating responses from query results.
Community responses on X highlighted enthusiasm for the practical tools. Dev Influencer Matt Pocock posted:
The AI SDK it’s the dominant AI lib in the TS ecosystem.
Meanwhile developer Divin Prince noted:
Been playing with vercel workflow write async/await code that can pause/resume, way simpler than managing queues yourself.
Overall, reactions praised the focus on production-ready features, with calls for more documentation on marketplace integrations. These beta tools are open for community testing, with Vercel inviting input to shape future stable versions.
