Skip to content
Depicta Depicta Docs

How to make images in Claude Code

Add the Depicta skill to Claude Code, and it can make the pictures your project needs: product photos, illustrations, icons, diagrams and slides, and it can edit or combine the images you already have.

Before you start

Section titled “Before you start”
  • A Depicta account. Sign up. You pay per image from prepaid credits, with no subscription. See Pricing.
  • Claude Code.

1. Add the skill

Section titled “1. Add the skill”
  1. In your dashboard, go to the Skills page.
  2. On the Depicta skill, choose Create & download.
  3. Name the skill after the computer, for example “my-laptop”. Give each computer its own name: downloading again with the same name replaces the earlier file and turns off its key.
  4. Under Format, choose Claude Code.
  5. Choose Create & download.
  6. Unzip the file into ~/.claude/skills/ (all projects) or <project>/.claude/skills/ (one project). The skill ends up at .../skills/depicta/SKILL.md.

The skill file includes its own API key, so there’s nothing else to set up.

2. Ask for an image

Section titled “2. Ask for an image”

Ask the way you’d brief a designer. Say what it’s for, the style, and any exact text.

Make a hero image for the landing page in this repo: a ceramic mug on a light oak table, soft morning light, 16:9.

Draw a flowchart of our sign-up flow from the routes in src/, flat style, our brand blue.

3. Use your own pictures

Section titled “3. Use your own pictures”

Point Claude Code at image files in your project to edit them, combine several, or match a style you already have.

Tips for better results

Section titled “Tips for better results”
  • Say what the picture is for. “Homepage hero”, “README banner” or “A0 poster” tells Claude the shape and how much detail it needs.
  • Ask for a set in one style. Ask for all the pictures in one request, or pass the first one as a reference, so they look like a set.
  • Put exact text in quotes, and end with “No other text”. For small text, keep it as real type in your page or document instead.
  • Edit in small steps. One change per request works better than a long list.
  • Ask Claude to finish the files: remove a background, crop to a size, or convert to JPEG or WebP. Those run in Depicta too.

What it costs

Section titled “What it costs”

Each image is charged from your credits at its actual cost, which depends on the model and the operation; our Pricing page shows the range for each model. Processing such as resizing or removing a background costs a fixed small amount per operation.

Good to know

Section titled “Good to know”
  • Every image Depicta generates carries Content Credentials (C2PA), which record that it is AI-generated, and an invisible watermark in each file we deliver, except very plain or transparent images and GIF or SVG files.
  • If you publish an image that could pass as a real photo of real people, places or events, say that it’s AI-generated. See Section 8.2 of the Terms.
  • Depicta is an independent service, not affiliated with or endorsed by Anthropic. Claude is a trademark of Anthropic.

Other ways to use Depicta

Section titled “Other ways to use Depicta”