How to Match Images to Your Slideshow Content in Shadarc
Prepare a relevant collection, add descriptions, and choose Match content in Advanced for Hook, Body, or CTA. Optionally exclude collection images with added text identified during analysis. Keep publishing manual, save the Automation, then pause its schedules and save again before checking examples in Library. Resume scheduled generation when you are ready.
Quick summary
To help your slideshow photos fit the words, prepare a relevant image collection, add descriptions, and choose Match content in your Automation. Pause its schedules and keep publishing manual while you generate and review examples, then decide whether to resume scheduled generation.
This guide uses a small desk-reset series as an example. The same approach can work for reading spaces, hobby supplies, home organization, or other subjects where you have several suitable photos to choose from.
For the thinking behind this approach, read how relevant images improve an automated slideshow.
1. Prepare a collection that supports your subject
In Media, create or open a collection for the series. Choose photos that give you several ways to illustrate the ideas you expect to cover.
For a desk-reset series, you might include an open work surface, a cable tray, a desk lamp, a notebook, and a small storage drawer. Keep the colors and overall feel reasonably consistent, while allowing enough variety to support different tips.
Start with clean photos that you are comfortable using in your posts. Remove unrelated subjects and images whose added captions would compete with your slide text.
A small, relevant collection can be easier to work with than a large collection mixing desks, food, travel, and unrelated quotes. Matching chooses from the collection you assign, so give it images that fit the series.
2. Analyze your photos or add descriptions yourself
For your own collection, choose Analyze images, then Start analysis. You can follow its progress and return to the collection afterward. If some images fail, use Retry failed analysis.
Analysis prepares descriptions for images that do not already have one. When you add more photos later, start analysis again to cover the new images.
You can also write descriptions manually. Open one of your photos, select the description below it, enter what the photo shows, and choose Save description.
You need saved descriptions to enable matching, but every photo does not have to be described before you begin. A more complete set of useful descriptions gives you more material to work with. Missing descriptions or unavailable matches can lead to Random selection from the collection.
If you use a Public collection, its descriptions are managed by Shadarc. You can read them, but editing and starting analysis are available for your own images.
3. Review descriptions for what is actually visible
Open a few photos and read their descriptions. Correct anything that would give the wrong impression of the subject.
Keep descriptions concrete. Name the main objects, what they are doing, and any detail that helps explain when the photo would be useful.
| Photo | Vague description | More useful description |
|---|---|---|
| A desk with an empty center | “Productive workspace” | “A wooden desk with a clear central work surface and a notebook at one side.” |
| Cables gathered in a tray | “Organized life” | “Charging cables gathered inside a shallow tray on a desk.” |
| A lit desk lamp | “Cozy evening” | “A desk lamp lighting a small work area beside a closed notebook.” |
Describe what you can see. Avoid inventing a person's routine, a product's benefits, or steps that the photograph does not demonstrate.
The description helps choose the image; it is not published on the slide. Analysis skips images that already have saved descriptions, so review and edit existing descriptions directly when they need improvement.
4. Give the Automation a brief your photos can support
Create an Automation from a Template that fits your series or a custom configuration, or open an Automation you already have. Then adjust its brief.
For the desk example, you could use:
Create a short slideshow for someone making a small desk easier to use after work. Open with a simple desk-reset idea, then offer three practical suggestions about a clear work surface, gathering cables, and lighting the work area. Use ordinary language. Avoid claims about health or productivity, and avoid describing the photos as proof that a particular routine works.
This is an example brief for an initial manual slideshow. For a recurring series, expand the range of suggestions only as your collection gains photos that can support them.
A useful brief and a useful collection work together. If your instructions call for a detailed drawing demonstration, a collection of decorative desk photos will give you little relevant material.
5. Choose Match content for the parts that need it
The Automation editor divides the slideshow into Hook, Body, and optional CTA. The Hook opens the post, the Body develops the idea, and the CTA asks the viewer to take an action.
For each part you want to configure:
- Select its tab and choose your collection in Media.
- Open Advanced.
- Under Image selection, choose Match content.
The collection and image-selection choices are separate for each part. You might use matching for the Body, where each tip needs a fitting photo, while leaving an atmospheric Hook on Random.
Before saving, open Delivery. Under Destination, choose Keep in Shadarc, or set Publishing mode to Manual if a TikTok account is selected. These Delivery settings apply to the whole Automation.
New Automations include an active daily schedule and need at least one active schedule for their first save. Leave a schedule active and choose Save changes once to create the Automation. Then open Delivery again, pause every schedule, and choose Save changes again before generating manual examples. For an existing Automation, you can pause its schedules and save directly.
The information icon beside Image selection shows description coverage and explains matching. If the collection has no saved descriptions, add descriptions or analyze it first. Choosing Match content does not start image analysis for you.
Matching uses the generated content to help choose images from the assigned collection. It can use Random selection when a suitable match is unavailable. A particular photograph is not guaranteed for a particular sentence.
6. Optionally exclude images with added text
In the same Advanced section, turn on Exclude images with text overlays if you want to skip photos with digitally added captions, quotations, or text watermarks identified during analysis.
This setting is separate for Hook, Body, and CTA, and works with either Random or Match content. Signs and labels naturally present in a photograph are allowed.
The setting relies on completed image analysis. Existing images are not automatically checked when you turn it on, and images with saved descriptions are skipped by analysis. Writing a description manually does not check a photo for added text. Review those photos yourself and keep clean originals in the collection when this matters to your series.
Check that the collection still contains enough suitable images after exclusions. The setting filters the available choices; it does not remove lettering from a photograph. If you enable Stock fallback, that option can supply images when a collection runs short; the overlay exclusion applies to collection images.
7. Generate manually and review the full slideshow
Choose Generate. The editor shows a Sent to Library message with a Library button. Open Library to follow generation and inspect the finished slideshow.
Read each slide's words beside its photo. For the desk example, a cable tip should have a photo that makes sense beside it, and an evening-lighting tip should not imply that an unrelated notebook illustrates lighting.
Check the image crop, text readability, and sequence as well as the individual pairings. The photographs should support the advice without making it look more specific or better evidenced than it is.
If you edit or regenerate the text, the slideshow keeps its selected images. Review those pairings again after changing the wording. Updated collection descriptions and Automation settings apply to future slideshows, so generate a new manual example to check those improvements.
8. Decide how much review the series needs
With schedules paused and publishing manual, try a few results with different content before setting the series to recur. Keep notes on repeated corrections, then improve the collection, descriptions, or brief that caused them.
In Delivery, adjust your schedule and choose Resume when you are ready for recurring generation. With Destination set to Keep in Shadarc, the output stays in Shadarc. If you select a TikTok account and set Publishing mode to Automatic, eligible slideshows publish after generation and validation without waiting for your individual approval. Choose Save changes to apply your updated schedule and publishing settings.
Automatic publishing also applies when you choose Generate manually. A scheduled start is not a promise of an exact posting minute.
Keep schedules paused and publishing manual if each post needs your sign-off. Our slideshow automation guide covers the broader setup and publishing choices.
For the next post, choose one clear subject and a collection that can show it well. Write or review the descriptions, enable matching where it helps, and judge the result from the words and images you actually see.
