Caveman Mode (Output Brevity)
Caveman mode makes Pando talk less. Think of the difference between a telegram and a letter: the news is the same, the paper is a tenth. Pando drops the greetings, the “let me explain what I am about to do” and the summaries of what it has just said, and keeps everything that matters.
It only changes how Pando expresses itself. It thinks, tests and checks exactly as much as before.
What it does for you
- Saves money. Models charge by the word they write, and Pando writes fewer.
- Saves reading time. The answer comes first, without a warm-up.
- Keeps long sessions lighter. Shorter answers fill the conversation more slowly.
How it feels in practice
There are three levels, like volume steps:
| Level | How Pando talks | Good for |
|---|---|---|
| Lite | Normal sentences without the padding. Answer first, then a short note | Everyday work |
| Full | Short lines and bullets, one idea each | Busy sessions where every word counts |
| Ultra | Only the bare facts | A tight budget or a very long conversation |
What gets cut: greetings and goodbyes, repeating your question back to you, announcements (“I’ll now…”), summaries of what was just said, apologies, praise and explanations nobody asked for.
What is never cut: code, commands and file paths, error messages, test results, web addresses, safety warnings, and anything you asked to hear in detail.
When to use it
Use it when you already know the ground and want results, not narration. Leave it off when you are learning something new and the explanations are the point.
Good to know
- It is off until you choose a level.
- You can pick a level for the current session with a command, and a different default for new sessions in the settings.
- A new session always starts with the default, not with what you chose in the previous one.
Next steps
- Guide: Change how Pando thinks and talks shows the commands and the setting.
- Guide: Spend fewer tokens covers the other ways to save.
- Reference: Working modes.