How Orbit agents use tools
Published Updated
This is the Learning Atlas from Orbit. Get your invite
Orbit-internal conceptsAI agents and harnessesConcepts and methodology
-
Step 1: Apollo
Why this matters here
Apollo starts the path because it is Orbit's concrete bridge between user intent, project context, and executable agent work.
In short
Orbit's built-in AI assistant. It runs in a floating window, knows your project state in real time, and handles operational tasks via natural language.
-
Step 2: MCP (Model Context Protocol)
Why this matters here
MCP shows the protocol boundary where Orbit-connected services become callable tools for an agent.
In short
An open standard protocol that enables AI models to securely connect to and interact with external custom tools, local filesystems, and enterprise data sources.
-
Step 3: Tool use
Why this matters here
Tool use is the visible behavior MCP enables, turning the agent from a text generator into a system that can inspect and act.
In short
The capability of an AI model to interact with external software, APIs, or scripts to perform actions it cannot do through text generation alone.
-
Step 4: Function calling
Why this matters here
Function calling explains the structured request shape behind many tool interactions, so learners can distinguish protocol from invocation.
In short
A specific type of tool use where an AI model is provided with a list of programming functions and intelligently generates the exact JSON arguments needed to execute them.
-
Step 5: Agentic loop
Why this matters here
After individual calls, the agentic loop shows how observation, decision, action, and revision repeat during real work.
In short
The continuous cycle where an AI agent observes its environment, makes a plan, executes a tool, evaluates the result, and decides what to do next.
-
Step 6: Token budget
Why this matters here
Token budget closes the path by showing the resource pressure that can end or distort long-running agent work.
In short
The strict allocation and monitoring of how many API tokens an AI agent or application is allowed to consume during a specific task or workflow.