Skip to main content

Documentation Index

Fetch the complete documentation index at: https://docs.toebox.ai/docs/llms.txt

Use this file to discover all available pages before exploring further.

ToeBox agents can use tools to interact with external systems, query data, and take actions on behalf of your customers. Tools are connected via the Model Context Protocol (MCP).

​
MCP tool connections

MCP (Model Context Protocol) tools connect your agents to live data and actions. ToeBox includes built-in connections to Shopify’s official MCP servers:

​
Shopify Catalog

Lets your agent search and browse products across your Shopify store. Capabilities:
  • Search products by keyword, category, or attribute
  • Retrieve product details (title, description, price, images, variants)
  • Filter by price range and availability
  • Browse collections
Endpoint: https://discover.shopifyapps.com/global/mcp

​
Shopify Checkout

Lets your agent create checkout sessions and guide customers through purchases. Capabilities:
  • Create checkout sessions
  • Collect buyer information
  • Attach payment methods
  • Complete purchases
Endpoint: https://{shop-domain}/api/ucp/mcp
MCP tools are auto-configured based on your agent type. For example, a Product Recommendation agent automatically enables the Catalog tool, while a Sales Assistant enables both Catalog and Checkout.

​
Data tools

When you connect data sources (Snowflake, PostgreSQL, or other databases), your agents gain access to built-in data tools:
ToolDescription
list_tablesLists all tables in the connected database
describe_tableReturns column names, types, and metadata for a specific table
query_databaseExecutes a read-only SQL query and returns results
These tools let agents answer data-driven questions like “What were our top-selling products last month?” by querying your data warehouse directly.
Data tools execute read-only queries. Write operations (INSERT, UPDATE, DELETE) are blocked at the connector level for safety.

​
Custom SQL tools

You can create custom SQL tools that wrap a specific query behind a friendly name and description. This is useful for:
  • Giving agents access to a frequently-used report
  • Restricting what data an agent can access
  • Pre-building complex queries so the agent does not need to generate SQL
To create a custom SQL tool:
  1. Go to your agent’s configuration
  2. Open the Custom Tools section
  3. Click Add Tool
  4. Provide a name, description, and SQL query
  5. The agent can then call this tool by name during conversations

​
Custom tool builder

For advanced use cases, the custom tool builder lets you define tools with:
  • Name — What the agent calls this tool
  • Description — Helps the AI understand when to use the tool
  • Parameters — Input parameters the tool accepts (with types and descriptions)
  • SQL query — The query to execute, with parameter placeholders
Write clear tool descriptions. The AI model uses the description to decide when to call a tool, so be specific about what it does and when it should be used.