Stitch::upload To Stitch

Upload-to-Stitch

Upload local assets (images, mockups, HTML, and markdown files) to a Stitch project using the provided upload script, which bypasses the MCP tool’s base64 output token limits.

[!NOTE] The AI model cannot upload files via MCP tools directly because the base64 encoding of even a small file exceeds the model’s output token limit (~16K tokens). This script reads the file and sends it directly over HTTP.

Steps

1. Identify Target Project

Use list_projects to find the correct projectId.

2. Get the API Key

Locate your active MCP server configuration file and extract the API key:

• Antigravity: .gemini/antigravity/mcp_config.json or .gemini/jetski/mcp_config.json
• Gemini CLI: ~/.gemini/settings.json or ~/.gemini/extensions/Stitch/gemini-extension.json
• Claude Code: ~/.claude.json

Extract:

• API Key: From the X-Goog-Api-Key header or auth argument
• MCP URL (optional): From the httpUrl or endpoint argument (defaults to https://stitch.googleapis.com)

[!IMPORTANT] If you cannot find the API key in any of these locations, or if you cannot access these files, you MUST ask the user to provide the Stitch API key. Do not proceed without a valid API key.

3. Run Upload Script

[!WARNING] Checkpoint — User Confirmation Required. Before running the upload script, you MUST pause and present the file(s) to be uploaded (paths, sizes, and types) to the user and wait for explicit approval. Do NOT execute the upload script until the user confirms.

Use run_command to execute the Python script:

python3 <SKILL_DIR>/scripts/upload_to_stitch.py \
  --project-id <PROJECT_ID> \
  --file-path <PATH_TO_FILE> \
  --api-key <API_KEY> \
  [--api-url <STITCH_API_URL>] \
  [--title <SCREEN_TITLE

[!TIP] macOS / SSL Certificate Troubleshooting: If the upload fails with ssl.SSLCertVerificationError: [SSL: CERTIFICATE_VERIFY_FAILED] unable to get local issuer certificate, this means your Python installation does not have root certificate authorities configured.

The script automatically attempts to use the certifi package to load the CA bundle if it is installed in your python environment. If certifi is not installed, you can either install it (pip install certifi) or manually supply the SSL_CERT_FILE environment variable when running the script:

bashCopy

SSL_CERT_FILE=$(python3 -c "import certifi; print(certifi.where())") python3 <SKILL_DIR>/scripts/upload_to_stitch.py \
  --project-id <PROJECT_ID> \
  --file-path <PATH_TO_FILE>

Supported File Types

The script auto-detects MIME type from the file extension.

Script Options

• --project-id: Required. The Stitch project ID.
• --file-path: Required. Path to the local file to upload.
• --api-key: Required. API key for Stitch authorization.
• --api-url: Optional. Base URL of the Stitch API. Defaults to https://stitch.googleapis.com.
• --title: Optional. Title for the uploaded screen. When uploading extracted HTML from a web app, set this to the route path of the page (e.g., '/dashboard', '/settings/profile', '/inbox') so that the screen name/title in Stitch clearly identifies the route.
• --generated-by: Optional. Specify how the uploaded file was generated (e.g., ‘stitch::extract-static-html’ skill, ‘Claude Code’, ‘Codex’, ‘Gemini’ etc.).