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

# Workbench

> Reusable Prompts, Personas, Knowledge, and Preview Context.

Workbench is the reusable context you keep in the Browser App: **Prompts** you insert into the composer, **Personas** that set the assistant's tone, **Knowledge** entries (labeled **Source** in the Library) that give a chat reference material, and **Preview Context**, which shows the assembled request before it goes out.

Included with **Advanced** and **Max**. Free and Plus accounts see the Workbench sections with a **View plans** action, and anything saved earlier stays in the browser after a downgrade.

Everything in Workbench is encrypted in your browser and stored locally. Nito receives it only inside the chat requests you send.

## Library and Active Context

Open the **+** control in the composer, or **Library & context** from the account menu.

* **Library** is where you create and edit Prompts, Personas, and Sources. A Source can be typed or built from a document; text-like files parse in your browser, larger documents such as PDFs go through Nito's no-store parser.
* **Active Context** is what applies to the current chat. Prompts insert once. Personas and Sources stay attached until you remove them.

Each Source remembers the privacy level it was created under and replays only into chats at that level or higher. The same rule governs [memory](/web-app/memory).

Chat summaries can be published into Knowledge from the summary review. Incognito chats never save to Knowledge.

## Prompts

Slash-style snippets for the jobs you repeat. Inserting one fills the composer; it does not persist as context.

## Personas

A Persona replaces the built-in Default Style with your own instructions, compiled into the system prompt under `User Persona`. New accounts start with **No persona**. One Persona is active per chat.

* **Settings > Text > Custom instructions** adds persistent guidance to every text chat, whatever Persona is active.
* **Tune Persona** on a reply turns your feedback into an updated Persona without copying the reply itself.

## Knowledge

Selected Sources are injected into the request under **Additional Context**, labeled as reference data rather than instructions. Keep each entry under the character limit shown in the editor.

## Preview Context

On Advanced and Max, **Preview Context** in the composer opens the exact request the send path will build: base prompt, active Persona, selected Sources, and an approximate token count for each. Free and Plus accounts find it described under **Settings > Intelligence**.

## Sharing

Prompts, Personas, and Sources share by encrypted link; the recipient imports a fresh local copy. See [Sharing](/web-app/sharing).

## Next

<CardGroup cols={2}>
  <Card title="Memory" icon="brain" href="/web-app/memory">
    Recall across chats.
  </Card>

  <Card title="Sharing" icon="link" href="/web-app/sharing">
    Encrypted links.
  </Card>

  <Card title="Backups and Sync" icon="cloud-arrow-up" href="/web-app/backups-and-sync">
    Workbench on every device.
  </Card>

  <Card title="Plans" icon="tags" href="/plans">
    Which plan includes Workbench.
  </Card>
</CardGroup>


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.