{
  "access": "public",
  "type": "tutorial",
  "format": "markdown",
  "title": "Sandbox Guide",
  "chunked": true,
  "url": "https://library.datagrout.ai/sandbox",
  "summary": "**Test individual tools directly in the browser**",
  "content_markdown": "# Sandbox\n\n**Test individual tools directly in the browser**\n\nEvery connected integration has a Sandbox tab where you can select a tool, fill in parameters, and execute it without an agent or SDK. The Sandbox shows the raw response, a generated cURL command, estimated credits, and optional chart previews.\n\n---\n\n## Getting Started\n\n1. Navigate to an integration's detail page\n2. Click the **Sandbox** tab\n3. Select a tool from the searchable dropdown\n4. Fill in parameters\n5. Click **Run**\n\nThe Sandbox is available on any integration that is connected to a server and has enabled tools. If no tools are enabled, go to the **Tools** tab first to enable them.\n\n---\n\n## Interface\n\n### Tool Selector\n\nA searchable dropdown listing all tools on this integration. Enabled tools appear first; disabled tools are shown below a divider. Each entry shows the tool name and description.\n\nType to filter the list, then click a tool to select it.\n\n### Parameters\n\nWhen you select a tool, its parameters populate automatically from the tool schema:\n\n- **Required parameters** are marked with an asterisk and cannot be removed\n- **Optional parameters** can be removed if not needed\n- Each parameter shows its name, type, and description\n- Enter values based on type: text for strings, JSON for objects and arrays, `true`/`false` for booleans\n\n**Global tool arguments** (Refract, Chart, Demux) are included based on your Interaction Settings. These can be toggled on or off per call.\n\nYou can also add **custom parameters** by entering a name and value in the fields at the bottom.\n\n### cURL Command\n\nThe Sandbox generates a live cURL command that mirrors your current configuration. It updates as you change parameters. Copy it to reproduce the call outside the browser.\n\n### Run Button\n\nClick **Run** to execute the tool. The button shows the estimated credit cost. If your balance is insufficient, the button is disabled with a tooltip showing the required credits.\n\n### Response\n\nThe right panel shows the result:\n\n- **Success**: Pretty-printed JSON response, including DataGrout metadata and receipt if enabled\n- **Error**: Structured error display with message, tool name, timestamp, and expandable technical details\n- **Chart**: If the tool returns chart data (or you enabled the Chart parameter), a chart preview appears above the response with a download button\n\n---\n\n## Use Cases\n\n- **Verify an integration works** after connecting it -- select a read-only tool and run it\n- **Test parameters** before writing agent code -- experiment with different arguments\n- **Debug a failing tool call** -- reproduce the exact call your agent made\n- **Generate cURL commands** -- copy the command for use in scripts or documentation\n\n---\n\n## Related\n\n- [Interaction Settings](interaction-settings) -- Configure global tool arguments that appear in the Sandbox\n- [Discovery Tools](discovery-tools) -- Find tools programmatically instead of browsing the dropdown\n- [Quick Start](quick-start) -- Other ways to test your server\n"
}