PromptLens — Image to Prompt

``` Every image is a prompt waiting to be read. PromptLens reverse-engineers any image on the web into a generation prompt — written in the native grammar of the model you're targeting, not a generic description. HOW IT WORKS 1. Right-click any image → "Generate prompt from image" 2. Pick a dialect — the result is cached, so switching targets is instant 3. Copy and paste into your generator For protected or canvas-rendered images, open the toolbar popup and drop, paste (Ctrl+V), or choose a file instead. SEVEN PROMPT DIALECTS • Midjourney — single-line v7 prompt with exact --ar computed from the source image and --stylize tuned to the look • Seedance 2.0 — treats the image as a first frame: Shot / Scene / Motion / Style structure for video generation • Kling / Veo — one flowing cinematography paragraph extending the still into motion • Flux / Nano Banana — dense natural prose, the way diffusion-transformer models respond best • SDXL / SD — weighted comma-separated tags plus a matched negative prompt • Album art brief — an art-director's cover concept • Detailed description — a maximally faithful, model-agnostic paragraph FULL-DETAIL MANDATE Every dialect enforces exhaustive coverage: every visible element, named colors ("oxidized copper green", not "greenish"), specific light direction and temperature, explicit composition. Filler adjectives are banned. The source's real aspect ratio is measured in your browser and carried into the prompt. BRING YOUR OWN KEY — TOTAL PRIVACY PromptLens has no backend, no accounts, no analytics. You use your own API key for Google Gemini, Anthropic Claude, OpenAI, or OpenRouter (which unlocks virtually any vision model via the custom model field). The key is stored only in chrome.storage.local on your device and is sent to exactly one place: the provider you chose. Images are converted to base64 locally and sent directly to that API. Nothing is logged, collected, or relayed. A free-tier Gemini key from Google AI Studio is enough for everyday use. OPEN SOURCE Full source code, landing page and privacy policy: https://github.com/YOUR_USERNAME/promptlens