Home Docs GitHub
apollo

Using apollo

The interactive loop, keybindings and commands.

The loop

Apollo AI Assistant
Type your prompt and press Enter to send. Type 'quit' to exit.
Available tools: ls, read, bash | Tab: autocomplete filenames

You: _

Type a prompt, press Enter, watch it think, execute tools and answer — rendered as formatted Markdown via glamour. Multi-turn context is kept, so follow-ups (“now refactor it”) just work.

Commands

Input Behavior
Regular prompts Sent to the model with conversation history
/ls, /read …, /bash …, /edit … Run a tool directly, no model call
/skill:<name> Load an agent skill, then send the rest of your message (Skills)
/usage Show session token usage (prompt / completion / total + request count)
quit Exit

Tab completion

Filenames autocomplete as you type — handy for read <Tab> and /bash <Tab> without leaving the prompt. Skill names complete after /skill: too.

Example session

You: list files in current directory
Thinking...
Executing [1/1]: ls
Thinking...
Apollo: Here's what I found...

You: read main.go
Thinking...
Executing [1/1]: read
Thinking...
Apollo: ...

You: quit