Connect Bloopy to your own APIs, databases, or CRMs to let your AI agent fetch and update records automatically.
The Toolkit feature allows you to connect your custom APIs, databases, CRMs, or external service endpoints directly to your Bloopy AI agent. Once published, your AI agent can automatically call these endpoints to perform real-time lookups (such as checking product stock) and execute transactions (such as reserving inventory) during customer conversations.
1. Uploading or Pasting Your Configuration
To configure your toolkit, navigate to Toolkit in the left sidebar. You can upload or paste a standardized JSON schema representing your tools:
- Download the Template: Click the 'Download the template' link at the top to see the required JSON schema structure.
- Upload JSON File: Drag and drop your `.json` configuration file, or click 'Select File' inside the upload zone to browse your local files.
- Paste JSON Content: Copy and paste your JSON schema directly into our interactive code editor under 'Paste JSON Content'.
2. Standardized JSON Schema Properties
Your JSON schema must define the toolkit metadata and an array of individual tools. Each tool object includes:
- toolkit_name: The descriptive title of your API integration (e.g., 'FinServe Banking API').
- name: The unique name of the specific tool (e.g., 'get_customer_balance'). Use lowercase letters and underscores.
- description: A clear explanation of what the tool does (e.g., 'Retrieve a customer's current account balance'). The AI uses this description to decide when to call the tool.
- route: The API endpoint route (e.g., '/api/v1/customers/{customer_id}/balance'). You can wrap path variables in curly braces.
- request_type: The HTTP request method (e.g., GET or POST).
- auth: The authentication schema (e.g., Bearer token or custom headers) used to secure your endpoint.
{
"toolkit_name": "FinServe Banking API",
"version": "1.0",
"tools": [
{
"name": "get_customer_balance",
"description": "Retrieve a customer's current account balance",
"route": "/api/v1/customers/{customer_id}/balance",
"request_type": "GET",
"auth": {
"type": "Bearer",
"token": "fs_live_8f2a9b3c7d5e1f4a6b8c2d9e0f1a3b5c"
}
}
]
}3. Validating and Publishing
After adding your schema, click the orange 'Validate JSON Configuration' button. Once validation succeeds:
- Fulfillment & Encryption: The configuration is validated for schema syntax and securely encrypted. Click 'Submit for Approval' or 'Publish' to deploy.
- Successful Status: You will see a 'Toolkit published successfully!' card showing the count of active tools available to your AI agent (e.g., '30 tools are now available').
- Download Current Config: Click 'Download Current Config' on the success card to save a copy of your active config file for backup.
4. Managing Active API Tools
Under Active Toolkit, you can review a paginated list of your active endpoints, HTTP methods (e.g. GET, POST), parameter counts, and descriptions:
- Replace Toolkit: To update your configuration, click 'Replace Toolkit' or 'Restart Config Flow' to run through the upload editor again.