Tool use
Published Updated
What is Tool use?
Tool use is an AI model’s ability to call external functions, APIs or programs, such as running a shell command or querying a database, to act beyond generating text.
The capability of an AI model to interact with external software, APIs, or scripts to perform actions it cannot do through text generation alone.
Tool use bridges the gap between an isolated AI brain and the real world. Instead of just replying with text, a model instructed to "use tools" will output a specific structured command (like a JSON payload). The surrounding software intercepts this payload, executes the requested action (like searching the web or reading a file), and returns the result back to the model.
How it relates to Orbit
Apollo and connected AI clients can call Orbit tools to read project state and perform approved actions. See MCP Tool Catalog for details.
When you would use it
You configure tool use when building applications where the AI must take physical digital actions, such as editing files or querying live databases, rather than just generating static text.
Common operations
- An AI searching the live internet to answer a query.
- An agent executing a bash script to compile code.
Frequently asked questions
Is tool use the same as function calling?
Mostly. Function calling is the API mechanism; tool use is the broader idea of a model choosing and calling tools, which MCP standardizes.
Which tools can AI clients use in Orbit?
Orbit exposes dozens of MCP tools, from reading project status to queueing prompts and running git, and Apollo calls them after confirmation. See MCP Tool Catalog.