Codia Studio

Overview

Codia Studio is Codia's AI design workspace. It combines an agent chat, image generation, visual search, image editing, and an editable canvas so you can move from an idea or reference image to usable design assets without switching between separate creative tools.

The Studio landing page describes the product as an AI design team: instead of choosing one fixed tool, you describe the outcome you need and work with Codia's design agents to generate, edit, refine, and export visuals.

Use Studio for:

• Brand, logo, poster, social, ad, packaging, print, presentation, UI, and e-commerce creative work
• Product shots, mockups, campaign visuals, thumbnails, icons, illustrations, and marketing assets
• Visual research, mood boards, and style references
• Image editing tasks such as background removal, upscaling, object cleanup, and visual refinement
• Rebuilding screenshots, images, or PDFs into editable design structure for downstream workflows

Where To Open It

• Open Studio workspace
• View the Codia Studio landing page

Use the landing page for product positioning, feature overview, examples, and pricing context. Use the Studio workspace for the actual creation flow.

Core Concepts

AI design agents

Studio supports multiple specialist design profiles. The default Codia agent is a generalist design partner, while specialist agents focus on clearer domains:

• Logo Designer
• Brand Designer
• Poster Designer
• Social Media Designer
• E-commerce Designer
• Amazon Designer
• Packaging Designer
• Ad Designer
• Illustration Designer
• UI Designer
• Print Designer
• Shopify Designer
• Temu Designer
• Presentation Designer
• Inspiration Finder

Choose a specialist when the target is clear. Use the default Codia agent when the task spans several categories or you are still exploring.

Prompt + references

Studio's prompt composer supports text instructions and image references. Reference images can be used in different modes:

Mode	Use it when
Auto	You want the agent to decide how to use the image
Edit	You want to modify the image directly
Style	You want a new output in the same visual style
Layout	You want to keep the composition or structure
Character	You want to preserve the subject or person

This is useful for workflows like "keep this layout, redesign it for a fintech brand" or "use this product photo, create a launch poster around it."

Generation controls

The Studio composer includes controls for:

• Model selection
• Inspiration or template style
• Aspect ratio
• Number of outputs
• Color constraints
• Example prompts

Keep these controls aligned with the final channel. For example, choose a vertical ratio for stories, a square ratio for app icons or social posts, and a landscape ratio for banners or presentation covers.

Common Workflows

Generate a new visual

1. Open the Studio workspace.
2. Choose the default Codia agent or a specialist designer.
3. Enter a concrete prompt with the asset type, audience, style, copy, and output size.
4. Set ratio, count, color, and inspiration options when needed.
5. Review the generated options.
6. Continue refining in chat or move the selected result into the canvas.

Example prompt:

Edit an existing image

1. Attach the source image.
2. Set the reference mode to Edit when you want to modify that exact image.
3. Describe the change: background, text, composition, object cleanup, product emphasis, or style.
4. Review the result and iterate with smaller follow-up changes.

Useful edits include:

• Remove or replace background
• Upscale a low-resolution image
• Clean up unwanted objects
• Remove watermark-like artifacts when appropriate
• Generate variations from the selected image
• Match a reference style or composition

Research visual direction

Use Inspiration Finder or visual search when you need references before creating final work. This is best for mood boards, style exploration, trend discovery, and competitive creative research.

After collecting references, select the strongest direction and generate new assets from it. Avoid mixing too many unrelated references in one prompt.

Convert visual assets into editable structure

Studio can connect to Codia's visual structuring capabilities. Use this when a source image, screenshot, or PDF is visually useful but too flat to edit directly. The generated structure can then be used for editing, design import, or downstream automation.

For API-oriented workflows, see Visual Struct API and PDF to Visual Struct.

Canvas Editing

Studio includes an editable canvas for continuing work after generation. The canvas and image action panels support workflows such as:

• Select, move, resize, and rotate elements
• Add text, shapes, frames, lines, and drawings
• Crop images
• Mask or select regions for chat-based image editing
• Erase objects or selected areas
• Expand image boundaries
• Generate variations from an image
• Insert generated assets into the canvas

Use the canvas when you need precise layout edits, composition cleanup, or manual finishing after an AI-generated result.

Library And Sessions

Studio work is organized around sessions. A session stores the conversation, generated artifacts, and associated canvas project. You can:

• Create a new session from the home composer
• Reopen recent sessions from the sidebar
• Rename or delete sessions
• Review generated artifacts
• Continue a previous conversation

Sessions are useful when a project needs multiple rounds of iteration or when you want to keep a record of prompt decisions and generated assets.

Output And Downstream Workflows

Studio output can feed other Codia workflows:

• Send UI-like assets to Screenshot to Design or Visual Struct workflows
• Use finished design frames as input for Design to Code
• Use presentation-oriented visuals with NoteSlide
• Export or reuse images, structured design data, SVG/PPTX/Figma-compatible assets where supported by the workflow

Studio is strongest when used as the creative and editing front door. For a narrow conversion task, use the dedicated product page or plugin instead.

Best Practices

• Name the asset type explicitly: poster, logo, product hero, app screen, slide cover, ad, banner, icon, etc.
• Include audience, channel, format, style, brand colors, required text, and non-goals.
• Use one or two reference images with an explicit reference mode.
• Iterate with focused follow-up prompts instead of rewriting everything.
• For product or brand work, review generated text, claims, prices, and logos before publishing.
• For UI work, convert complex screens section by section.
• Use specialist agents when you know the design category.

Troubleshooting

The result is visually good but not on brief

Make the prompt more concrete. Add the channel, audience, required copy, visual hierarchy, style constraints, and what to avoid.

The reference image is being used the wrong way

Change the reference mode. Use Edit for direct edits, Style for look and feel, Layout for composition, and Character when the subject must stay consistent.

The output misses details

Use a higher-quality input, crop to the relevant area, or split dense layouts into smaller tasks. Long pages, packed dashboards, and text-heavy images usually work better in sections.

The generated text is inaccurate

AI-generated and OCR-derived text should be checked manually before release, especially numbers, names, claims, prices, and legal copy.

The canvas feels easier than chat for the next step

Move into canvas editing when the next change is spatial or precise. Use chat when the next change is conceptual, stylistic, or generative.