> ## Documentation Index
> Fetch the complete documentation index at: https://docs.budecosystem.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Quick Start

> Create and run your first agent in Bud AI Foundry

This guide walks through creating a basic project-scoped agent, validating it, and preparing it for reuse.

```mermaid theme={null}
graph TD
    A[Open Project Agents Tab] --> B[Click Create New]
    B --> C[Select Agent Type]
    C --> D[Choose Model or Deployment]
    D --> E[Add Prompt and Variables]
    E --> F[Run Agent]
    F --> G[Save]
```

## Step 1: Open the Agents Surface

1. Sign in to Bud AI Foundry.
2. Navigate to a project.
3. Open the **Agents** tab.
4. Click **+ Create New**.

## Step 2: Start the Workflow

1. Confirm project context.
2. Choose an agent type.
3. Continue to prompt configuration.

## Step 3: Configure Prompt Behavior

1. Select the target model/deployment.
2. Add a system prompt.
3. Add user/assistant messages as needed.
4. Optionally enable structured input/output and define variables.

## Step 4: (Optional) Connect Tools

1. Open the **Tools** panel.
2. Search and select a connector.
3. Complete credential flow (client credentials or authorization code).
4. Confirm connector appears in the connected list.

## Step 5: Run and Save

1. Click **Run Agent** to validate behavior.
2. Refine prompts/settings if needed.
3. Save the agent.

## Step 6: Reuse the Agent

1. In the project table, click **Use this agent**.
2. Copy cURL, Python, or JavaScript snippets.

<Tip>
  After initial validation, create a new version for controlled updates instead of overwriting production behavior.
</Tip>