I've been working on a Persian classical poetry archive with LLM-generated translation. As a poetry enthusiast but a non-speaker of Persian, I wanted to read Rumi, Hafiz, and Bedil, but found myself jumping between ChatGPT and Google Translate. I sourced my data from Ganjoor, which was luckily available via their public API. Then I quickly put together a website with bring-your-own-key translation feature.
For now, there's no backend and it's hosted on GitHub Pages; it's a static website reading chunks from an sqlite-based DB file. I intend to change it in the future by running an LLM (preferably a frontier Gemini model, as works great for Persian) over the whole data set so you won't need a key to read the translated poetry. However, I don't have the budget of $1000-$2000 to spend on the tokens currently.
> so you won't need a key to read the translated poetry
This probably sounds like a crazy question, but if you are already getting translations using someone else's API keys, can't you just store/cache that response so that the next time the translation is requested, you already have it?
get a few personal gmail accounts's gemini API key, and you have 1000s of free LLM calls to flash-lite models, which are more than upto this task. one time run, and save it to your db.
For now, there's no backend and it's hosted on GitHub Pages; it's a static website reading chunks from an sqlite-based DB file. I intend to change it in the future by running an LLM (preferably a frontier Gemini model, as works great for Persian) over the whole data set so you won't need a key to read the translated poetry. However, I don't have the budget of $1000-$2000 to spend on the tokens currently.
Website: https://0x5ce.github.io/farsidaan
Code: https://github.com/0x5CE/farsidaan
P.S. Contributions welcome.