What is ai_melli?

ai_melli is a Persian-language project that turns a GitHub repository into a free AI image and text generator, run entirely through GitHub Actions rather than a hosted server. There is no API key to obtain and no separate signup: everything runs on GitHub's own infrastructure, using the repository owner's Actions minutes. The project offers three workflows. Image generation takes a prompt, a negative prompt for things to exclude, a model choice such as flux, an image width and height, a seed, and how many images to produce, then returns AI-generated pictures. Text generation takes a prompt, a model choice such as an OpenAI-style model, a temperature setting, and an optional system message, and returns AI-written text for things like articles, stories, or code explanations. Smart image generation accepts a prompt in any language, including Persian, uses a text model to first translate and improve the prompt, then feeds the improved version into the image model, which the README says produces noticeably better results than typing a raw prompt directly. After each run, results are collected into a simple HTML gallery page with a dark theme that includes the original prompt, the improved prompt if one was used, the model, the date, and all generated images. Output can be committed back into the repository or attached to a GitHub Release for a permanent, easy to share link. Watermarks are removed from generated images. Setup involves forking the repository, then enabling GitHub Actions with read and write workflow permissions under the repository's settings, which the README stresses is safe to do on your own fork. From there, you trigger any of the three workflows manually from the Actions tab, fill in the parameters you want, and run it. This tool is aimed at Persian-speaking users who want to generate AI images and text without paying for an API key, using GitHub's free Actions compute as the entire backend.