Chrome MV3 extension that generates WCAG-compliant alt text for images using local Gemma 4 E2B or cloud AI providers
TypeScript
0
45 commits
updated May 5, 2026
A Chrome MV3 extension that generates WCAG-compliant alt text for any image via right-click context menu, using a three-tier inference cascade:
npm install
npm run build
chrome://extensionsdist/ folderThe extension is now active. Right-click any image on any page and choose Generate Alt Text.
npm run dev # watch mode — rebuilds on save
npm run typecheck # TypeScript check (both DOM and worker contexts)
npm test # unit tests (Vitest, 31 tests)
npx playwright test # integration tests (requires built dist/)
After editing source files in watch mode, go to chrome://extensions and click the refresh icon on the extension card to reload it in the browser.
Open the extension's options page (click the extension icon → Options, or visit chrome://extensions → Details → Extension options) to configure:
chrome.storage.local (device-only, never synced){style} as a placeholdersrc/
types.ts # Shared types across all extension contexts
storage.ts # chrome.storage abstraction (sync for settings, local for keys)
prompt.ts # Prompt construction + {style} interpolation
inference/
chrome-ai.ts # Chrome Built-in AI (Prompt API) provider
cloud.ts # Anthropic / OpenAI / Gemini / Custom providers
cascade.ts # Tier selection logic
background/
service-worker.ts # Context menu, cascade routing, offscreen lifecycle
offscreen/
offscreen.html # Offscreen document entry point
offscreen.ts # Gemma 4 E2B host (WebGPU, promise-cached model)
content/
content.ts # Injected script: tracks right-click, handles messages
popup.ts # Shadow DOM popup component
options/
options.html / .css / .ts
tests/
unit/ # Vitest unit tests
integration/ # Playwright extension tests
45 commits
TypeScript
84.7%
CSS
7.5%
HTML
6.2%
JavaScript
1.6%
Chrome MV3 extension that generates WCAG-compliant alt text for images using local Gemma 4 E2B or cloud AI providers
TypeScript
0
45 commits
updated May 5, 2026
A Chrome MV3 extension that generates WCAG-compliant alt text for any image via right-click context menu, using a three-tier inference cascade:
npm install
npm run build
chrome://extensionsdist/ folderThe extension is now active. Right-click any image on any page and choose Generate Alt Text.
npm run dev # watch mode — rebuilds on save
npm run typecheck # TypeScript check (both DOM and worker contexts)
npm test # unit tests (Vitest, 31 tests)
npx playwright test # integration tests (requires built dist/)
After editing source files in watch mode, go to chrome://extensions and click the refresh icon on the extension card to reload it in the browser.
Open the extension's options page (click the extension icon → Options, or visit chrome://extensions → Details → Extension options) to configure:
chrome.storage.local (device-only, never synced){style} as a placeholdersrc/
types.ts # Shared types across all extension contexts
storage.ts # chrome.storage abstraction (sync for settings, local for keys)
prompt.ts # Prompt construction + {style} interpolation
inference/
chrome-ai.ts # Chrome Built-in AI (Prompt API) provider
cloud.ts # Anthropic / OpenAI / Gemini / Custom providers
cascade.ts # Tier selection logic
background/
service-worker.ts # Context menu, cascade routing, offscreen lifecycle
offscreen/
offscreen.html # Offscreen document entry point
offscreen.ts # Gemma 4 E2B host (WebGPU, promise-cached model)
content/
content.ts # Injected script: tracks right-click, handles messages
popup.ts # Shadow DOM popup component
options/
options.html / .css / .ts
tests/
unit/ # Vitest unit tests
integration/ # Playwright extension tests
45 commits
TypeScript
84.7%
CSS
7.5%
HTML
6.2%
JavaScript
1.6%