Help & FAQ

Use the search box on the homepage, browse by category or model, or filter results by model, type, category and language on the search page. You can also browse tags from the menu.

On a prompt page, use “Copy full prompt” to copy everything (including parameters like --ar), or “Copy positive” for the prompt text only. On mobile, tap a card in the feed to open it, then tap “Copy Prompt”.

TXT (plain text), Markdown (formatted, good for Notion/Obsidian) and JSON (structured, for developers/tools). Downloads contain the prompt text, parameters, source and license — never images.

No. Tap ☆ on any card or prompt page to save it locally — you can add notes, copy in batch and export from My favorites. Creating a free account syncs your favorites across devices automatically.

Sample images are hot-linked from their original sources (we never store image files). If a source site is down, rate-limits or removes the image, it may not load — the prompt text itself is always available. Broken images are periodically checked and pruned.

Prompts are written for specific AI image/video generators, and each model has its own syntax. Each model page explains its writing style and parameters. “通用/General” means the source didn't specify a model.

Prompts are collected for learning and reference — check each prompt's license field and the source page. Generated images' usage rights depend on the AI model you run them through, not on us.

Use the “Report” form at the bottom of any prompt page. Copyright issues are handled with priority — we remove infringing content once verified.