> ## Documentation Index
> Fetch the complete documentation index at: https://docs.trysidebar.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# System Prompt Preview

## What it does

System Prompt Preview shows you the complete system prompt that sideBar sends to your
AI provider at the start of a conversation. This includes everything injected
automatically: your profile, memories, open content, skill definitions, and
any other context sideBar has assembled. It also shows a breakdown of how much
of your provider's context window is being used and by what.

It is a transparency and debugging tool. If sideBar is behaving unexpectedly,
a system prompt preview can help you understand what it knows and what instructions it
is working from.

## Actions

| Action                         | Description                                                                                                                                     |
| ------------------------------ | ----------------------------------------------------------------------------------------------------------------------------------------------- |
| Generate system prompt preview | Assembles and displays the full system prompt for the current conversation, with a section-by-section breakdown and context window usage report |

## When sideBar uses it

Only when you ask for it explicitly:

* "Show me your system prompt"
* "What instructions are you running under?"
* "What context are you working with right now?"
* "Preview the prompt"

## Requirements

* **Tier:** Standard (included, no installation needed)
* **API key:** None required

## How to enable

System Prompt Preview is enabled by default. To toggle it, go to **Settings → Skills**.

## Opening the preview

When the preview is ready, sideBar shows a button in the conversation to open
it. The skills pill confirms that the skill ran, but the preview itself opens
from the message button.

## What the preview shows

The preview includes:

* **System prompt**: the full text sent to your AI provider at the start of
  every message
* **Section breakdown**: which parts of the prompt come from where (your
  profile, memories, open content, skill definitions, and so on) with
  character counts for each
* **Context window report**: total characters used versus the limit for your
  current provider and model, and whether any content was trimmed to fit

## Examples

> "Show me your system prompt."

> "What do you know about me right now?"

> "How much of your context window are we using?"