LLLucas Lewisinapimodels.hashnode.dev·13h ago · 6 min readWe Replaced a 2,300-Character Image Prompt With One SentenceOur prompt for making AI images look real used to be 2,300 characters long. It listed every AI tell to remove, every camera detail to add, and rules for colour, shadows and lens optics. It produced fa00
LLLucas Lewisinapimodels.hashnode.dev·19h ago · 25 min readThe 6 Best Image Text Editors in 2026: An In-Depth Guide to Changing the Words Already in a PictureThe poster has gone to print, the promo graphic is scheduled, and one word on it is wrong. The designer who made it is away, the layered source file is on their laptop, and all you have is a flat JPEG00
LLLucas Lewisinapimodels.hashnode.dev·19h ago · 6 min readHow to Read C2PA and IPTC AI Labels From Image Bytes in Node.jsMost AI images carry a note from the tool that made them, and you can read it with about 60 lines of TypeScript. The hard part is not parsing it; it is knowing what its absence means, which is nothing01M
LLLucas Lewisinapimodels.hashnode.dev·19h ago · 14 min readOne Upscaler Can't Do Anime and Photos: Routing Images by Type in Next.jsThe first time we put a generative HD model behind our image upscaler, it turned anime characters into real people. A flat-shaded anime girl went in and a photorealistic woman came out, same pose, new00
LLLucas Lewisinapimodels.hashnode.dev·4d ago · 10 min readHEIC uploads in the browser: decode natively first, fetch the 2.9 MB WASM decoder only when you mustBrowser-side image preparation is the step between a user picking a file and the upload starting, where the file is decoded, converted and shrunk so the server never sees a format or a size it cannot 00
