Connecting Actions in Your Application

Defining an action in the Studio (details, inputs, response mock, and evaluations) completes the setup inside Agent Studio.

The final step is for your front-end developer to implement the action in your application code so the agent can actually execute it.

Why This Step Matters

• In the Studio, you define what the action does and confirm the AI can route prompts and extract schema fields.
• In your application, a developer connects the action to real logic (API calls, database updates, navigation, etc.).

What to Share with Developers

When handing off, provide your developer with:

• Action Key → unique identifier of the action (e.g., create_task, invite_user).
• Schema → the input fields defined in the Studio (e.g., email, organizationRoleType, name).
• Expected Output → what the agent should perform and if there's any relevant data the agent should leverage in the chat response like deep-linking to a page

Developers will use these to wire the action into your product’s code.

Using MCP (Model Context Protocol)

If your developers are using AI-powered coding tools like Cursor, VS Code, or Windsurf, the defined actions and schema are available via MCP. This lets them pull your Studio-defined actions directly into their coding environment, speeding up implementation.

Next Steps

• PMs / AI leads: finalize and Publish the action version you want live.
• Developers: connect the published action in code.

📘 See the Developer Guide → Coding Actions for full implementation details. ⚡ See Vibe Coding with MCP for setup in Cursor, VS Code, or Windsurf.

👩‍💻

Think of the Studio as the blueprint for the action, and your application code as the engine that makes it run.