NeutronField

NeutronField publishes prompt guides for the generators currently in use and explains what each finished image carries. Each guide covers a generator or technique, how the model reads a prompt, the settings that affect results, and worked examples the editors ran themselves. It is written for people who create images with these tools, including designers, hobbyists, marketers, and illustrators, and notes that there are no accounts and no sales on this domain.

The editors run prompts themselves on the model named beside the guide, and they read each file exactly as the generator handed it over, before resizing or conversion. They validate content credentials with an open-source library and report what the vendor documents say.

It describes several core file-carried elements: a content credential that signs the tool and records edits; an invisible watermark detectable by vendor detectors; metadata fields that indicate software writes like source type; and a prompt saved in the file that records front-end prompts. It also explains that missing a label does not prove human authorship of an image and provides guidance on how prompts are written and evaluated across models.

The site documents the tool’s plans and scope, including integrations with OpenAI and Gemini APIs and references to other vendors and model families. It frames its content as a publication of prompt guides and file-record explanations, with sections on prompts for different generators, how the guides are written, and how to read file content and provenance.

Use cases

• For designers building mood boards and hobbyists running open models
• For marketers, illustrators, and educators