Images
A great blog post deserves a great cover image. Wonderblogs generates AI images that match your brand, and learns your preferences over time.
Generating images
The Generate tab offers two ways to create images: pick a pre-built template, or use the full pipeline with a custom prompt.
Templates
Templates are pre-built prompt structures where you fill in a few parameters (headline, icons, colors, mood) and Wonderblogs assembles the generation prompt for you. They skip the content analysis pipeline and go straight to image generation, so they're fast and predictable.
Ten templates are available:
- Hand-Drawn Sketch (16:9): marker-style black ink illustration with a single accent color. Good for editorial blog headers.
- Blog Hero Photo (16:9): photorealistic editorial photograph with precise lighting and film stock emulation.
- Infographic Diagram (16:9): clean infographic-style illustration with data points and a bold headline.
- Abstract Background (16:9): abstract art background suitable for overlaying text. Customizable colors and mood.
- Social Media Card (1:1): square-format image optimized for Instagram and LinkedIn.
- YouTube Thumbnail (16:9): bold, eye-catching image with clear focal point and space for text overlay.
- Instagram Story (9:16): vertical mobile-optimized image for Instagram Stories and TikTok.
- Hero Banner (21:9): ultra-wide banner with gradient colors and a text-safe zone for website hero sections.
- Product Shot (4:3): clean e-commerce product photography with customizable surface and lighting.
- Portrait Photo (2:3): editorial portrait with precise lighting control and optional film stock emulation.
To use a template, select it in the grid, fill in the required fields, and click Generate Image. The global image model set by the admin applies to all template generation. If your site has brand colors configured, they are automatically incorporated into the generated image for visual consistency.
Custom prompts
Without selecting a template, describe what you want in the prompt field and click Generate Image. This runs the full 5-stage pipeline: content distillation, visual brief, prompt synthesis, generation, and evaluation. The pipeline uses your brand's visual identity (color palette, style, mood) from Brand Context to produce on-brand visuals.
Generation typically takes 30–60 seconds with real-time progress updates. The finished image is saved to your site's gallery automatically.
Gallery
The Gallery tab shows every image associated with your site:
- Standalone images: created from the Generate tab.
- Post cover images: generated during content runs.
- API images: generated via the Image Generation API.
Each image shows a source label (Standalone, Post, or API) on hover so you can tell where it came from. Click any image to see its generation prompt and rating status. Standalone images can be deleted from the gallery; post cover images and API images are managed from their respective pages.
Teaching Wonderblogs your style
The Feedback tab shows unrated images for you to review. Rating them is how Wonderblogs learns what you like.
For each image you rate five dimensions on a 1–5 scale:
- Brand Alignment: Does it match your brand identity?
- Content Relevance: Is it relevant to the topic?
- Visual Style: Does the style match your preferences?
- Mood & Tone: Is the emotional quality right?
- Composition: Is the layout and visual quality good?
You can also tag specific issues (e.g., "Has text," "Too abstract") and leave written feedback. After 5 or more ratings, Wonderblogs summarizes your preferences and uses them to guide future image generation, so your visuals get better over time.
Reference Images
For the most precise visual control, upload reference images in the Visual Identity section of your Brand Context page. These are example images that represent your ideal cover image style. Wonderblogs analyzes them to extract colors, style, composition, and mood, then uses those insights to guide every image it generates.
To add reference images:
- Go to Brand Context and scroll to the Visual Identity section.
- Drag and drop images into the Reference Images area, or click to browse. JPG, PNG, and WebP are supported (max 5 MB each).
- Each image is automatically analyzed by an AI vision model. You'll see a checkmark when analysis is complete.
Free plans allow up to 2 reference images; Pro plans allow up to 5. You can view the extracted style notes by expanding the Extracted Style Notes panel below your uploaded images.
Reference images work alongside the existing visual identity text fields (color palette, style, mood) and your feedback ratings. Think of them as "show, don't tell": the AI sees exactly what you mean instead of interpreting text descriptions.
API Jobs
The API Jobs tab shows image generation jobs triggered via the Image Generation API (POST /api/v1/images/generate). Each job displays its status, webhook delivery state, and timestamps.
Click any row to expand it and see additional details: the generation prompt, visual brief, result image, or error message if the job failed.
While jobs are pending or processing, the table refreshes automatically every few seconds. Completed API images also appear in the Gallery tab alongside your other images.
Visual Identity settings
The Generate tab shows a summary of your current visual identity. To update your color palette, style, mood, or banned visual elements, head to the Brand Context page and edit the Visual Identity section.