> For the complete documentation index, see [llms.txt](https://help.usejimo.com/docs/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://help.usejimo.com/docs/experiences/agent/knowledge/interface.md).

# Interface

Let the Agent read your own product's pages so it can answer questions about what users actually see on screen, element by element.

***

**Dashboard path:** [Agent > Knowledge > Interface](https://i.usejimo.com/agent/knowledge)

Sources teach the Agent what your documentation says. Interface teaches it what your product **is**: the screens, buttons, filters and menus your users are looking at when they ask for help.

<figure><img src="https://2794860263-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FKzAcDWQbK1gKbpra7bkb%2Fuploads%2FBLmShcvWW8M1BeLv6QDD%2F04-00_interface-tab-empty-state.png?alt=media&amp;token=e042a8d3-6474-4010-9ab1-942bfa78c0dd" alt=""><figcaption></figcaption></figure>

## Two ways the Agent can learn a page

When you add a page, Jimo asks: [**How should your agent learn this page?**](#user-content-fn-1)[^1]

| Mode                | What it does                                                                       | Time     | Control    |
| ------------------- | ---------------------------------------------------------------------------------- | -------- | ---------- |
| **Scan and curate** | Catalogs the elements on the page. You review everything before the Agent uses it. | \~3 mins | You curate |
| **Read all**        | The Agent reads the page in real time whenever a user asks about it. No setup.     | Instant  | Sees live  |

<figure><img src="https://2794860263-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FKzAcDWQbK1gKbpra7bkb%2Fuploads%2Fsdthpw9nVUk3NHhdPFE5%2F07_two-learning-modes.svg?alt=media&amp;token=02383c00-857a-45f8-9c89-53bff392f14c" alt=""><figcaption></figcaption></figure>

**Scan and curate** is the right choice for your core screens: you keep or hide each element found by the scan, write what each one does, and attach Skills to the page. The Agent sees only what you kept, plus form fields and open dialogs; an element hidden here can never be the target of a Guide.&#x20;

**Read all** fits long-tail pages where instant coverage matters more than curation: no element list is stored, the Agent reads the whole live page (still with the privacy masking) whenever the eye is on.

{% hint style="info" %}
**Neither mode decides whether the Agent answers in text or guides on screen.** That is the Skill's type and the user's eye. The mode decides what a running Skill can see and point at: Read all gives the widest coverage, Scan and curate gives the most control. On a page that is not scanned at all, buttons and links are invisible to the Agent, so a Guide has nothing to point at.
{% endhint %}

#### Enabling Read all

Read all is the same thing as turning **Trusted Page** on for that page: the Agent gets full live access and can read any element on the fly, without a catalog. Because of that, Jimo asks you to confirm you understand before enabling it, then shows a confirmation screen with two ways to continue: **Test Chat**, or **Set up skills**.

<figure><img src="https://2794860263-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FKzAcDWQbK1gKbpra7bkb%2Fuploads%2FIACLdjCn5TXCOLnIdIBc%2F04-04_read-all-confirmed.png?alt=media&amp;token=92b2d4fd-0d79-4a92-8dbb-16bb0b550a76" alt=""><figcaption></figcaption></figure>

{% hint style="warning" %}

#### **Read all means the Agent sees what the user sees, live.**&#x20;

If a page can display sensitive data, prefer Scan and curate and turn off the elements that should stay private. See [Page details](/docs/experiences/agent/knowledge/interface/page-details.md).
{% endhint %}

You can switch a page from one mode to the other at any time, from the **Trusted Page** toggle in [Page details](/docs/experiences/agent/knowledge/interface/page-details.md).

### The pages list

Each scanned page appears as a card with its screenshot, name, and two counters: the number of **Elements** the Agent can read, and the number of **Skills** attached to the page.

A page that has not been rescanned for 30 days is flagged to the Agent as possibly outdated. Rescan after any UI change on a page that carries Skills.

<figure><img src="https://2794860263-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FKzAcDWQbK1gKbpra7bkb%2Fuploads%2Ff4NGkHbmpgDimjQ0hMli%2Fimage.png?alt=media&amp;token=2a4a3a5d-fb9d-43bd-bc3b-294b788995f0" alt=""><figcaption></figcaption></figure>

Use the **search** field to find a page, and **Scan a page** (top-right) to add one. See [Scan a page](/docs/experiences/agent/knowledge/interface/scan-a-page.md). There is no limit on the number of pages you can add.

Clicking a card opens the page's detail panel: its elements by category, its Skills, and its URL rules. See [Page details](/docs/experiences/agent/knowledge/interface/page-details.md).

### In this section

* [**Scan a page**](/docs/experiences/agent/knowledge/interface/scan-a-page.md): capture a page's elements, interact to uncover more, and save
* [**Page details**](/docs/experiences/agent/knowledge/interface/page-details.md): curate what the Agent reads, element by element

### Related pages

* [Knowledge](/docs/experiences/agent/knowledge.md): the four kinds of Agent knowledge
* [Skills](/docs/experiences/agent/skills.md): procedures that run on the pages the Agent knows
* [Sources](/docs/experiences/agent/knowledge/sources.md): the documentation side of the knowledge base

[^1]: <img src="https://2794860263-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FKzAcDWQbK1gKbpra7bkb%2Fuploads%2FHmPuEg3cgaHJcTOAYstM%2F04-02_how-should-agent-learn-page.png?alt=media&amp;token=df50c01e-eab8-4dad-b311-2c6ed99761bd" alt="" data-size="original">