Vercel Ai Sdk Best Practices Skill

• Use streamText for streaming text responses from AI models.
• Use streamObject for streaming structured JSON responses.
• Implement proper error handling with onFinish callback.
• Use onChunk for real-time UI updates during streaming.
• Prefer server-side streaming for better performance and security.
• Use smoothStream for smoother streaming experiences.
• Implement proper loading states for AI responses.
• Use useChat for client-side chat interfaces when needed.
• Use useCompletion for client-side text completion interfaces.
• Handle rate limiting and quota management appropriately.
• Implement proper authentication and authorization for AI endpoints.
• Use environment variables for API keys and sensitive configuration.
• Cache AI responses when appropriate to reduce costs.
• Implement proper logging for debugging and monitoring.

Iron Laws

1. ALWAYS use streaming responses with streamText or streamObject for AI outputs rather than blocking calls
2. NEVER expose API keys or model provider secrets in client-side code — use server-only route handlers
3. ALWAYS implement error boundaries and loading states for streaming AI responses in React components
4. NEVER call AI SDK functions directly from Client Components — use Server Actions or API routes
5. ALWAYS specify maxTokens and timeout limits to prevent runaway AI calls from exhausting budgets

Anti-Patterns

Memory Protocol (MANDATORY)

Before starting:

cat .claude/context/memory/learnings.md

After completing: Record any new patterns or exceptions discovered.

ASSUME INTERRUPTION: Your context may reset. If it's not in memory, it didn't happen.