PromptLens

PromptLens turns any on-page image — photo, product shot, UI capture — into structured creative intelligence across three modes: • Style Prompt — ready-to-paste prompts for Midjourney, Stable Diffusion, Flux, and DALL·E / GPT-4o, with --ar, --style, and weighting parameters pre-filled. • Product Visual — six-dimension analysis of any product photo: Color, CMF, Lighting + Lens, Composition + Scene, Mood + Props, and a reusable generation Prompt. • Web Design — extract typography, palette, components, interactions, and DTCG-compliant design tokens from any page. Export as Skill.md (for AI agents) or Tokens Studio JSON (for Figma). HOW IT WORKS 1. Hit the shortcut (default ⌘⇧Y on Mac, Ctrl+Shift+Y on Windows/Linux). 2. Drag to select any region on the active page. 3. Pick a mode in the side panel, get structured output in seconds. 4. Copy, export to skill.md, or paste straight into your creative tool. BRING YOUR OWN KEY PromptLens ships without a server. Plug in OpenAI, Anthropic, Google Gemini, or any OpenAI-compatible endpoint (Azure, self-hosted, regional clouds). The request goes from your browser directly to your chosen provider, using your own API key. Nothing is sent to us. Today PromptLens is 100% local-first and BYOK. A managed service tier is on our roadmap — the BYOK option will always remain. PRIVACY • No accounts, no telemetry, no backend. • API key lives in chrome.storage.sync (synced only to your own Google account). • Images stay in memory; never stored or sent anywhere other than your configured endpoint. • Last 50 generations cached locally in IndexedDB; clear any time. PERMISSIONS (MINIMUM) • activeTab — inject the selection overlay on the page you triggered • storage — remember your API config and local history Built for designers, brand researchers, product teams, and anyone compounding a visual vocabulary from scrolling Pinterest, Dribbble, Xiaohongshu, X, or their own competitor pages. PromptLens is local-first and open about what it does.