Back to Copilotkit

Reading agent state

showcase/shell-docs/src/content/docs/integrations/pydantic-ai/shared-state/in-app-agent-read.mdx

1.70.15.8 KB
Original Source

<IframeSwitcher id="shared-state-read-example" exampleUrl="https://feature-viewer.copilotkit.ai/pydantic-ai/feature/shared_state?sidebar=false&chatDefaultOpen=false" codeUrl="https://feature-viewer.copilotkit.ai/pydantic-ai/feature/shared_state?view=code&sidebar=false&codeLayout=tabs" exampleLabel="Demo" codeLabel="Code" height="700px" />

<Callout type="info"> This example demonstrates reading from shared state in the [CopilotKit Feature Viewer](https://feature-viewer.copilotkit.ai/pydantic-ai/feature/shared_state). </Callout>

What is this?

You can easily use the realtime agent state not only in the chat UI, but also in the native application UX.

When should I use this?

You can use this when you want to provide the user with feedback about your agent's state. As your agent's state updates, you can reflect these updates natively in your application.

Implementation

<Steps> <Step> ### Run and connect your agent <RunAndConnect components={props.components} /> </Step> <Step> ### Define the Agent State
Create your Pydantic AI agent with a stateful structure. Here's a complete example that tracks language:

```python title="agent.py"
from textwrap import dedent

from pydantic import BaseModel
from pydantic_ai import Agent, RunContext
from pydantic_ai.ui import StateDeps
from pydantic_ai.ui.ag_ui import AGUIAdapter
from starlette.applications import Starlette
from starlette.requests import Request
from starlette.responses import Response
from starlette.routing import Route


class AgentState(BaseModel):
    """State for the agent."""
    language: str = "english"


agent = Agent("openai:gpt-5.4-mini", deps_type=StateDeps[AgentState])


@agent.instructions()
async def language_instructions(ctx: RunContext[StateDeps[AgentState]]) -> str:
    """Instructions for the language tracking agent.

    Args:
        ctx: The run context containing language state information.

    Returns:
        Instructions string for the language tracking agent.
    """
    return dedent(
        f"""
        You are a helpful assistant for tracking the language.

        IMPORTANT:
        - ALWAYS use the lower case for the language
        - ALWAYS response in the current language: {ctx.deps.state.language}
        """
    )


async def run_agent(request: Request) -> Response:
    # Build the deps fresh on every request: `dispatch_request` writes the state
    # the client sent into `deps.state`, so a shared instance leaks state between
    # concurrent requests and users.
    return await AGUIAdapter.dispatch_request(
        request, agent=agent, deps=StateDeps(AgentState())
    )


app = Starlette(routes=[Route("/", run_agent, methods=["POST"])])

if __name__ == "__main__":
    import uvicorn
    uvicorn.run(app, host="0.0.0.0", port=8000)
```
</Step> <Step> ### Use the `useAgent` Hook With your agent connected and running, call the `useAgent` hook, wait for the real agent, and initialize any missing UI-owned state with `agent.setState`.
```tsx title="ui/app/page.tsx"
import { useEffect } from "react";
import { useAgent } from "@copilotkit/react-core/v2"; // [!code highlight]

// Define the agent state type, should match the actual state of your agent
type AgentState = {
  language: "english" | "spanish";
}

function YourMainContent() {
  const { agent, isReady } = useAgent({
    agentId: "my_agent",
  });
  const state = (agent.state ?? {}) as Partial<AgentState>;

  useEffect(() => {
    if (!isReady || state.language !== undefined) return;
    agent.setState({ ...(agent.state ?? {}), language: "english" });
  }, [agent, isReady, state.language]);

  // ...

  return (
    // style excluded for brevity
    <div>
      <h1>Your main content</h1>
      <p>Language: {state.language}</p>
    </div>
  );
}
```

<Callout type="warn" title="Important">
  The `name` parameter must exactly match the agent name you defined in your CopilotRuntime configuration (e.g., `my_agent` from the quickstart).
</Callout>

<Callout type="info">
  The `agent.state` in `useAgent` is reactive and will automatically update when the agent's state changes.
</Callout>
</Step> <Step> ### Give it a try! As the agent state updates, your `state` variable will automatically update with it! In this case, you'll see the language set to "english" after the connected agent is ready. </Step> </Steps>

Rendering agent state in your app

You can render the agent state in any React component under <CopilotKit>. Read agent.state and return ordinary JSX.

tsx
import { useAgent } from "@copilotkit/react-core/v2"; // [!code highlight]

// Define the agent state type, should match the actual state of your agent
type AgentState = {
  language: "english" | "spanish";
};

function YourMainContent() {
  const { agent } = useAgent({
    agentId: "my_agent",
  });
  const state = (agent.state ?? {}) as Partial<AgentState>;

  if (!state.language) return null;
  return <div>Language: {state.language}</div>;
}
<Callout type="warn" title="Important"> The `name` parameter must exactly match the agent name you defined in your CopilotRuntime configuration (e.g., `my_agent` from the quickstart). </Callout> <Callout type="info"> The `agent.state` in `useAgent` is reactive and will automatically update when the agent's state changes. </Callout>

Intermediately Stream and Render Agent State

By default, the Pydantic AI Agent state will only update between Pydantic AI Agent node transitions -- which means state updates will be discontinuous and delayed.