Image Studio, Generate Images, Infographics & Carousels

The Image Studio is the home for visual generation in FeedBoss. From a single prompt, you can produce an image, an infographic, or a carousel, choose the format, pick a style or template, attach reference images, and generate. Every result is editable in place with AI-powered actions like remove object, change background, edit text in image, and more.

Open it from the workspace sidebar (Studio) or directly at /workspace/[id]/studio.

What you can make

The format chip in the prompt bar switches the output type:

• Image, a single still image, generated freely or from a curated image template (quote card, stylized vector, single stat, illustration, and more).
• Infographic, a single visual that explains a concept at a glance. Picks up your selected style preset or one of the infographic templates.
• Carousel, a multi-slide story for the LinkedIn feed. Generated via streaming and dropped into your Recents when ready, with the editor one click away.

The Generate hero

The hero is the prompt bar at the top of the Studio. It accepts a full description of what you want to create (press / anywhere on the page to jump to it).

Around the prompt are five chips that shape the output:

• Format, Image, Infographic, or Carousel.
• Aspect ratio, Square (1:1), Portrait (2:3), Landscape (3:2), or Widescreen (16:9). Carousel locks to square; infographic defaults to portrait. Widescreen is available on the Agency plan.
• Style, pick a curated visual style. Each style controls colors, typography, and overall aesthetic so the output looks consistent without manual styling.
• Reference, attach up to 4 images total to guide the output. Each reference is tagged with a role:
  • Style (up to 2), copy the look, color palette, and typography of the reference.
  • Subject (up to 2), keep the subject/person/object recognizable in the output.
  • Logo (up to 1), place your logo in the result (best for infographics).
• Refine, rewrite your brief into a polished, model-ready prompt. The refined prompt bakes in your selected style; once refined, the chip stays “refined” until you change the prompt or pick a different style (then it offers Re-refine).

Press Cmd/Ctrl + Enter in the prompt to generate.

Brand Kit

If your workspace has a Brand Kit, the Studio shows a Brand Kit toggle near generation controls. Turning it on applies the saved handle, tagline, logo, typography, and brand colors to the generation request. The toggle uses the Brand Kit’s primary color as a small status dot when enabled.

The same Brand Kit preference is available in post-attached image creation, image templates, infographic generation, carousel generation, and image edits, so brand styling can follow the asset from first draft through refinement.

Prompt length

The Studio gives you room to write a real brief:

• Image and Infographic prompts: up to 4,000 characters
• Carousel prompts: up to 6,000 characters

A live counter sits under the prompt input and turns red as you approach the cap. The Generate button is disabled once you go over, so you’ll never silently submit a truncated prompt. The same limits apply to the Create Image modal attached to posts and to all template input forms.

LinkedIn “AI-generated” label

Images generated through FeedBoss are processed before they reach LinkedIn so the platform no longer flags them with its automatic AI-generated label. This applies to both OpenAI and Gemini-generated images, including edits performed in the studio.

Templates

Below the hero, the Templates explorer lists ready-made starting points organized by format:

• Image templates are grouped by category (quote card, stylized vector, single stat, illustration, …). Picking one opens a focused input form where you fill in the content the template needs.
• Infographic and Carousel templates work the same way, pick a template, fill in the content, and generate.

Some templates are marked Pro. Trial and Starter accounts see an upsell tile in each row pointing to the full library on Pro/Agency.

A preview modal lets you inspect any template before committing, click a card to open the preview, then Use this template to load it into the input form.

Recents

The Recent section shows everything you’ve generated in the workspace, images, infographics, and carousels, mixed together and sorted newest first. From any recent tile you can:

• Regenerate, drops the original prompt back into the hero so you can tweak and re-run it.
• Use as reference, adds the recent image to your reference list (tagged as a style reference by default).
• Add to post, attach it to one of your posts.

Edited images carry their parent’s lineage, so re-runs from history start from the cleaned-up prompt rather than the internal [action_id] {…} payload.

Result card & AI editing

When generation finishes (image and infographic), the Result Card takes over the screen. Carousel results jump straight to the Canvas editor.

From the Result Card you can:

• Regenerate, re-run the same prompt + style combination.
• Edit prompt, go back to the hero with the prompt preserved.
• Use as reference, feed the result back as a reference for the next generation.
• Add to post, attach the result to a post.
• Edit with AI, open the Edit Panel (see below).

Edit Panel

The Edit Panel exposes nine AI editing actions. Each action knows which inputs it needs (text params, an optional reference image, an optional brush mask):

Action	What it does
Add object	Insert a new object or element into the image.
Remove object	Erase a selected object cleanly. Pairs well with the Element Removal flow.
Replace object	Swap one object for another.
Change lighting / weather	Restyle the scene’s lighting or weather.
Change background	Replace the background while keeping the subject.
Apply style	Restyle the entire image (e.g. adopt a reference’s style).
Edit text in image	Change text rendered inside the image.
Translate text in image	Translate in-image text into another language.
Freeform	Anything else, describe the edit in plain language.

For spatial actions (anything that targets a region), turn on the Brush mask to paint exactly where the edit should apply. The brush toolbar supports size adjustment, undo, redo, and clear.

For Remove object specifically, the studio also offers an Element Removal flow: it auto-detects objects in the image (people, text blocks, decorations, etc.), lets you tap up to ten to select, and removes them in one pass. Detection results are cached, so re-opening the panel is near-instant.

After an edit completes, a Before / After slider is available on the result so you can drag to compare the original and the edited image.

Keyboard shortcuts

• /, focus the prompt from anywhere on the Studio page.
• Cmd/Ctrl + Enter, submit the prompt.

Tips

• Style references and a Style preset are mutually exclusive on carousel. If you attach a style reference to a carousel generation, the analyzed style takes over and the selected style preset is ignored.
• Refine is most useful on rough briefs. A one-line idea (“post about deep work”) expands into a detailed, render-ready prompt. The pulse on the Refine button after you pause typing is a hint that refinement would help.
• Carousel generation streams. You’ll see live progress (e.g. 3/8 slides) in the loading card; when it finishes the carousel lands in Recents and is ready to open in the Canvas editor.