What the Generator Does
You write a description, and the tool draws it as a pencil sketch. That is the whole workflow: text goes in, one new image comes out. Nothing is traced or filtered, so the drawing is created from scratch every time.
Behind the Generate button there are two steps:
- The description is rewritten. A text model turns what you typed into a detailed English prompt that asks for pencil qualities: monochrome graphite tones, clear linework, soft shading, cross-hatching and visible paper texture. It also steers away from words that would make the result look like a photograph.
- The image is rendered. An image model draws that prompt, and the result appears under the form, ready to download.
Because of the rewrite step, you can describe the scene in any language. "Un vieux phare sur une falaise" and "an old lighthouse on a cliff" lead to the same kind of drawing.
The tool works from words only. It does not accept an uploaded photo and cannot turn an existing picture into a sketch. If you want a drawing of a particular scene, describe it in as much detail as you can.