Dynamic Tools

A Tool is an action an AI client can perform — calling an API endpoint, running a database query, or executing a chain. Tools are what turn a passive API into something an AI agent can actually act on.

What you can do

• Generate tools automatically from every operation in an imported OpenAPI/Postman spec.
• Add, edit, or delete tools by hand when you need full control over what's exposed.
• Define the JSON Schema input and output for every tool, so the AI client knows exactly what to send and what it will get back.
• Test a tool directly from the app before handing it to an AI client.
• Attach notes and comments to a tool so teammates know why it exists or what changed.
• Turn a tool's raw JSON response into a rendered HTML view using a Handlebars template, for AI clients that can display rich output. See How to Add HTML Output to a Tool.

How to use it

1. Open a server and go to its Tools tab.
2. Select a tool to review or edit its name, description, and schema — clear, specific wording here directly affects how well an AI client uses it.
3. Use the built-in test panel to call the tool and inspect the response before relying on it.
4. Enable or disable individual tools to control what's visible to AI clients and on the public documentation page, without deleting them.

Good to know

• A tool's description is what the AI reads to decide when and how to call it — treat it like documentation aimed at a very literal reader.
• Related: Chains for combining several tools into one call, MCP Server Management.