Picking a Runtime

Choosing the right runtime is crucial for your assistant-ui implementation. This guide helps you navigate the options based on your specific needs.

Quick Decision Tree

Core Runtimes

These are the foundational runtimes that power assistant-ui:

Pre-Built Integrations

For popular frameworks, we provide ready-to-use integrations built on top of our core runtimes:

Understanding Runtime Architecture

How Pre-Built Integrations Work

The pre-built integrations (AI SDK, LangGraph, etc.) are not separate runtime types. They're convenient wrappers built on top of our core runtimes:

• AI SDK Integration → Built on LocalRuntime with streaming adapter
• LangGraph Runtime → Built on LocalRuntime with graph execution adapter
• LangServe Runtime → Built on LocalRuntime with LangServe client adapter
• Mastra Runtime → Built on LocalRuntime with workflow adapter

This means you get all the benefits of LocalRuntime (automatic state management, built-in features) with zero configuration for your specific framework.

When to Use Pre-Built vs Core Runtimes

Use a pre-built integration when:

• You're already using that framework
• You want the fastest possible setup
• The integration covers your needs

Use a core runtime when:

• You have a custom backend
• You need features not exposed by the integration
• You want full control over the implementation

Feature Comparison

Core Runtime Capabilities

Available Adapters

Common Implementation Patterns

Vercel AI SDK with Streaming

import { useChatRuntime } from "@assistant-ui/react-ai-sdk";

export function MyAssistant() {
  const runtime = useChatRuntime({
    api: "/api/chat",
  });

  return (
    <AssistantRuntimeProvider runtime={runtime}>
      <Thread />
    </AssistantRuntimeProvider>
  );
}

Custom Backend with LocalRuntime

import { useLocalRuntime } from "@assistant-ui/react";

const runtime = useLocalRuntime({
  async run({ messages, abortSignal }) {
    const response = await fetch("/api/chat", {
      method: "POST",
      headers: { "Content-Type": "application/json" },
      body: JSON.stringify({ messages }),
      signal: abortSignal,
    });
    return response.json();
  },
});

Redux Integration with ExternalStoreRuntime

import { useExternalStoreRuntime } from "@assistant-ui/react";

const messages = useSelector(selectMessages);
const dispatch = useDispatch();

const runtime = useExternalStoreRuntime({
  messages,
  onNew: async (message) => {
    dispatch(addUserMessage(message));
    const response = await api.chat(message);
    dispatch(addAssistantMessage(response));
  },
  setMessages: (messages) => dispatch(setMessages(messages)),
  onEdit: async (message) => dispatch(editMessage(message)),
  onReload: async (parentId) => dispatch(reloadMessage(parentId)),
});

Examples

Explore our implementation examples:

• AI SDK v6 Example - Vercel AI SDK with useChatRuntime
• External Store Example - ExternalStoreRuntime with custom state
• Assistant Cloud Example - Multi-thread with cloud persistence
• LangGraph Example - Agent workflows

Common Pitfalls to Avoid

LocalRuntime Pitfalls

• Forgetting the adapter: LocalRuntime requires a ChatModelAdapter - it won't work without one
• Not handling errors: Always handle API errors in your adapter's run function
• Missing abort signal: Pass abortSignal to your fetch calls for proper cancellation

ExternalStoreRuntime Pitfalls

• Mutating state: Always create new arrays/objects when updating messages
• Missing handlers: Each UI feature requires its corresponding handler (e.g., no edit button without onEdit)
• Forgetting optimistic updates: Set isRunning to true for loading states

General Pitfalls

• Wrong integration level: Don't use LocalRuntime if you already have Vercel AI SDK - use the AI SDK integration instead
• Over-engineering: Start with pre-built integrations before building custom solutions
• Ignoring TypeScript: The types will guide you to the correct implementation

Next Steps

1. Choose your runtime based on the decision tree above
2. Follow the specific guide:
   • AI SDK Integration
   • LocalRuntime Guide
   • ExternalStoreRuntime Guide
   • LangGraph Integration
3. Start with an example from our examples repository
4. Add features progressively using adapters
5. Consider Assistant Cloud for production persistence