Source: https://docs.pageloop.ai/guides/correct-images-in-drafts

# Get correct images into drafts

# Get correct images into drafts

A successful capture is not the same as a correct documentation image. Check that each screenshot shows the screen, account view, and state described by the surrounding instructions. Use automated retakes when Pageloop can reach that view, or capture the screen yourself when it cannot.

> [!NOTE]
>
> These steps apply to work reviewed in **Inbox**. An Admin manages app logins under **Account** → **Settings** → **Capture** → **Automated screenshots**. An Admin or Editor can answer a **Screenshots needed** request. Contributors do not see **Ask** in the sidebar.

## Decide which screenshots need replacing

Treat a screenshot as stale when the product no longer matches the image: a control moved or changed its label, a field was added or removed, or the workflow reaches a different screen. Also replace a capture of the wrong environment, account role, or example state. Age alone does not tell you whether an image is correct.

Before retaking images, identify the affected pages and the exact screens those pages need. Use representative sample data and check for personal or sensitive information. Do not use a sign-in page, access error, or unrelated screen as a substitute for the requested image.

## Request automated retakes

Use automated capture when the saved app account can sign in and reach the intended screen. If access or sign-in is the problem, have an Admin correct the login or the app account's permissions first. Do not weaken a necessary capture restriction to obtain a screenshot.

### Refresh several screenshots together

1. Open **Ask** and describe the pages and screenshots you want refreshed. List the affected pages or sections rather than asking for an unspecified set of images.
2. Explain what changed and which app environment and account view the images should show. For example: “Refresh the screenshots in the workspace setup guides to show the revised team menu. Keep images of screens that still match.”
3. Review the resulting proposed changes in **Inbox**. Check every replacement in context; requesting several retakes does not mean every capture will succeed or be included in a draft.

### Retake one image from an Inbox item

1. Open the item and select the affected page from the left-hand navigation.
2. Open **Ask** from the page header. Identify the image by its page, section, and what it shows, and ask Pageloop to retake it. Include the state the replacement must show.
3. Inspect the resulting proposal and confirm that the intended image changed. Do not assume that starting a retake has already replaced the image in the draft you are reviewing.

> [!WARNING]
>
> Ask cannot retake a screenshot for a brand-new, unpublished page in an Inbox item. Capture the image yourself and replace it in the **Article** tab instead.

## Resolve a Screenshots needed request

Pageloop can finish a proposal without obtaining every screenshot. Open the item's **Activity** tab and look for **Screenshots needed**. The request names the missing screens and groups them by the reason automation could not capture them.

| Reason shown by the request                    | What to do                                                                                                                                                 |
| ---------------------------------------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------- |
| Automated screenshots are not set up           | Have an Admin set up an app login, or supply the requested images manually.                                                                                |
| Pageloop could not sign in                     | Have an Admin check the saved login and sign-in instructions.                                                                                              |
| The screenshot account cannot open the screen  | Give the app account the appropriate access, or use an account that can capture the screen manually.                                                       |
| The app address is wrong                       | Have an Admin check the screenshot login's app address.                                                                                                    |
| The screenshot browser stopped partway through | Use **Run screenshots again** when the request offers it. This retries the item's screenshot work together rather than uploading an image for each screen. |

The settings links on the request lead to the relevant login setup. A failed capture does not always mean the password is wrong; follow the reason shown.

### Capture and upload the requested screen yourself

1. Read the requested screen's name. In your app, sign in with an account that can reach that screen and open the exact view the article describes.
2. Prepare safe example data and capture the product interface with your usual screenshot tool. Exclude unrelated desktop content and personal or sensitive data.
3. Return to **Screenshots needed**. Beside the matching screen, select **Upload a screenshot** and choose the image. Use PNG, JPEG, or WebP, no larger than 10 MB.
4. Check the uploaded preview and filename against the requested screen. Use **Replace** if you chose the wrong image.
5. Repeat for the other screens you can supply, then select **Place these screenshots**. Wait for **Checking and placing your screenshots…** to finish.

Uploading a file alone does not place the image. **Place these screenshots** starts a follow-up run that checks the supplied images for safety and whether they match the requested screens, then places suitable images in the proposed documentation. Review the resulting draft before continuing to publication.

You can submit some screens without answering every row; unanswered screens remain requested. Use **Skip** for a screen you deliberately want to leave without a supplied screenshot, then submit with **Place these screenshots**. Skipping does not create an image.

### Capture with the Chrome extension

Instead of uploading a saved file, select **Capture with the Chrome extension** beside the requested screen. Install the extension if prompted, then use the recording to reach the screen and take the intended screenshot. Finish the recording and return to the Inbox item.

The request uses the last captured image from the recording, so capture one requested screen at a time. Check the returned preview, then select **Place these screenshots**. The extension capture still needs the same placement and review steps as a file upload.

## Review images inside an Inbox item

### Compare the before and after image

1. Open the item's **Changes** tab and expand the image change.
2. Compare **Old** with **New**. **No image** means that side has no image, such as an addition with no previous screenshot. **Image preview unavailable** means the preview failed to load; it does not mean the image was deliberately removed.
3. Select the affected page and use **Article** to check the image beside the instructions it illustrates. Confirm the screen, labels, permissions, example data, and framing.
4. **Accept** a correct proposal or ignore an incorrect change. If the replacement is still wrong, request a retake or add a manual capture before finishing the item.

![Inbox image change comparing Old, with No image, against a proposed New screenshot](/images/uploads/screenshot-64ea8afd.png)

### Choose automated or manual capture

| Situation                                                                               | Use                                                                                                              |
| --------------------------------------------------------------------------------------- | ---------------------------------------------------------------------------------------------------------------- |
| An existing published page needs a fresh view that the saved app account can reach      | Ask for an automated retake, as described above.                                                                 |
| Automation cannot reach the screen, or the image needs a state you can prepare yourself | Capture the image manually. Answer **Screenshots needed** through its matching row when that request is present. |
| The page is brand-new and unpublished                                                   | Capture and replace the image manually in **Article**.                                                           |
| The right image is already available on the item                                        | Open **Media**, inspect the image, and drag it into the article where it belongs.                                |

**Media** groups images by **Pageloop**, **Extension**, and **Uploaded**. Search or filter the drawer, click a thumbnail to inspect it, then drag the image into the article. The check mark indicates that an image is already in that article; it is not approval of the image's accuracy.

Adding an image directly in the article and answering **Screenshots needed** serve different purposes. Use the request card to send a missing capture to the follow-up run. Use the article editor to control the finished draft's placement. Attaching an image to a comment is not a substitute for either step.

## Confirm the images reached the draft

After an automated retake or a screenshot-placement run finishes, reopen the affected page and check the resulting image, not just the capture count or an uploaded thumbnail. Confirm that the draft no longer shows the stale image and that the replacement agrees with the surrounding steps.

Wait for manual edits and review decisions to save. Capturing, uploading, placing, or accepting an image does not publish the documentation. Finish the item's normal review and publishing process, then check the image on the published page.

## Related

- [The Pageloop sidebar](/guides/pageloop-sidebar) — find Ask, Inbox, and screenshot settings.
- [Review an Inbox item](/guides/review-inbox-item) — review changes, answer questions, and collaborate in comments.
