MeiGen launches weekly prompt library with model filters and open dataset export

• MeiGen launched as a searchable library of AI image prompts pulled from trending X posts, pitched as a way to stop good prompt recipes from disappearing into likes and bookmarks launch thread.
• The library is updated weekly and, according to site overview, focuses on hand-picked prompts with real engagement rather than random uploads.
• MeiGen's feature list says users can filter prompts by model, including Nano Banana Pro, GPT Image, and Midjourney, then generate from a prompt in one click.
• The dataset is also open to fork and reuse; open dataset says it includes the trending prompts and their associated data via the linked GitHub repo GitHub repo.

MeiGen is a prompt archive for creators who already source ideas from X but do not want to manually save and sort every useful post. In the launch thread, the project is framed as a weekly scrape-and-curate workflow: popular prompt posts are collected, organized, and made searchable instead of getting buried in social feeds.

The initial product pitch is practical rather than model-specific. MeiGen's feature list highlights weekly updates, model filters, one-click generation, saved collections, and visible like and view counts so users can judge which prompts are actually getting traction.

The more interesting part is the export layer. The project's open dataset says the full dataset is open source and can be used, forked, or built on, which turns MeiGen from a browsing tool into a reusable prompt corpus for studios, researchers, or anyone building their own reference library. The public site lives at MeiGen site.

That matters because prompt libraries usually break at the organization step. MeiGen's demo clip shows the basic pain point clearly: good prompt posts are easy to discover and hard to retrieve later. By adding model filters and an exportable dataset, MeiGen is packaging prompt hunting as a repeatable archive instead of a bookmarks habit.