Custom Toolkits

The Custom Toolkits ability lets you attach custom tools you've built to your AI workflows. This extends the AI's capabilities beyond the built-in abilities with your own integrations, API calls, and business logic.

Settings

How It Works

Custom Toolkits connect your AI to external services and custom logic that you define. When this ability is enabled, the AI can invoke the tools in the selected toolkits during conversations.

Toolkits are collections of one or more tools. Each tool defines:

• A name and description (so the AI knows when to use it)
• Input parameters
• The action to perform (API call, webhook, etc.)

For full details on creating and managing toolkits, see the Toolkits developer guide.

Adding Custom Toolkits

Custom Toolkits can be added at either the workflow or task level:

1. Open your workflow or task
2. Go to the Abilities tab
3. Click Add Ability and select Custom Toolkits
4. Select the toolkits to include
5. Save

Task-level abilities override workflow-level abilities of the same type.

Tips

• Write clear tool descriptions so the AI knows when each tool is appropriate to use.
• Test tools independently before attaching them to production workflows.
• Start with one toolkit and add more as needed to keep behavior predictable.
• Healthcare workflows: For EHR integration, scheduling, lab results, and clinical decision support examples, see Healthcare MCP Integration Examples.

Conditional Activation

You can restrict when Custom Toolkits are available using CEL expressions in the ability's Conditions field. For example, you might only enable certain toolkits for specific member labels. See Abilities for more on conditional abilities.

Related

• Abilities — Overview of all AI abilities
• Toolkits — Build custom tools and MCP integrations
• Workflows — Attach toolkits to workflow tasks