Toolkits

Toolkits are collections of custom tools for your AI. Build your own tools or connect external services - then attach them to workflows via the Custom Toolkits ability.

Finding Toolkits

Go to Toolkits to see:

• All Toolkits - Built-in and custom toolkits available
• Custom Toolkits - Your workspace's custom tools
• MCP Servers - External integrations

Creating a Custom Toolkit

1. Go to Toolkits → Custom Toolkits tab
2. Click Create Toolkit
3. Set name, slug, and description
4. Add tools to the toolkit

Creating Custom Tools

Each tool needs:

Input/Output Schemas

Define what your tool accepts and returns using JSON Schema:

json

Data Access

Grant tools access to:

• Files - Read from or write to specific files/folders
• Data Types - Read from or create form records

Using Toolkits

Attach toolkits to workflows via the Custom Toolkits ability:

1. Open a workflow
2. Go to Abilities tab
3. Add Custom Toolkits ability
4. Select which toolkits to include

Built-in Toolkits

Your workspace may have built-in toolkits based on enabled features:

MCP Servers

Connect external services via MCP (Model Context Protocol):

1. Go to Toolkits → MCP Servers tab
2. Add server connection
3. Authenticate if required
4. Tools from the server become available

Example: Order Lookup Tool

Toolkit: "Customer Service"

Tool: check_order_status

• Prompt: "Look up the order and return its current status and tracking information"
• Data Access: "Orders" form (READ)
• Input: { orderId: string }
• Output: { status: string, tracking: string }

The AI can now check order status during conversations.

Healthcare Examples

For healthcare-specific toolkit and MCP integration examples, including EHR queries, appointment scheduling, lab results, drug interaction checking, and clinical decision support, see Healthcare MCP Integration Examples.