From stitch-kit
Uploads screenshots, mockups, or wireframes as new screens to Stitch projects. Enables redesign workflows by importing UI images for editing or code conversion.
How this skill is triggered — by the user, by Claude, or both
Slash command
/stitch-kit:stitch-mcp-upload-screens-from-imagesThis skill is limited to the following tools:
The summary Claude sees in its skill listing — used to decide when to auto-load this skill
Uploads one or more images (screenshots, mockups, wireframes) into a Stitch project as new screens. This is the entry point for the "redesign existing UI" workflow — import what you have, then use `edit_screens` to iterate or convert directly to code.
Uploads one or more images (screenshots, mockups, wireframes) into a Stitch project as new screens. This is the entry point for the "redesign existing UI" workflow — import what you have, then use edit_screens to iterate or convert directly to code.
Only use this skill when the user explicitly mentions "Stitch".
You must have a projectId before calling this. If you don't have one:
stitch-mcp-create-projectstitch-mcp-list-projectsUse the helper script to convert a local image file to base64:
bash scripts/encode-image.sh "path/to/screenshot.png"
The script outputs raw base64 to stdout. Capture it for the API call.
Supported formats and their MIME types:
| Extension | MIME type |
|---|---|
.png | image/png |
.jpg, .jpeg | image/jpeg |
.webp | image/webp |
.gif | image/gif |
{
"name": "upload_screens_from_images",
"arguments": {
"projectId": "3780309359108792857",
"images": [
{
"fileContentBase64": "[base64-encoded-image-data]",
"mimeType": "image/png"
}
]
}
}
projectId — numeric ID only, no prefix✅ "3780309359108792857"
❌ "projects/3780309359108792857"
images — array of image objectsEach image needs:
| Field | Type | Description |
|---|---|---|
fileContentBase64 | string | Base64-encoded image data (no data: prefix) |
mimeType | string | MIME type matching the image format |
You can upload multiple images in a single call — each becomes a separate screen.
Returns session info similar to generate_screen_from_text. The uploaded images appear as new screens in the project.
stitch-mcp-list-screens to find the new screen IDsstitch-mcp-edit-screensstitch-mcp-get-screen → framework conversionstitch-mcp-generate-variantsscripts/encode-image.sh — Base64 encoding helpernpx claudepluginhub gabelul/stitch-kit --plugin stitch-kitGenerates UI screens in Google Stitch from markdown prompt files using MCP tools. Parses sections by markers, sends prompts for generation, fetches images and code, saves to exports and code folders.
Creates Stitch UI screens, design systems, variants, and prototypes from briefs, mockups, or product context. Exports Tailwind HTML, screenshots, and local code artifacts. Use for screen exploration, editing Stitch projects, or multi-screen prototypes.
Orchestrates Stitch UI design workflows: routes from user request through spec, prompt assembly, screen generation, iteration, design system extraction, and framework conversion (Next.js, Svelte, HTML, React Native, SwiftUI).