Dev vibecodes 150-line FB2 converter to feed books to LLMs, ditching ad-riddled tools for pure markdown bliss

A developer has created a custom FB2 converter for large language models, allowing users to convert FB2 e-books into structured Markdown or plain text. The tool, built using Python and Streamlit, features a web UI and command-line interface, enabling users to upload files, read books directly in the browser, and download the result in their desired format. The converter preserves the semantic structure of the narrative, mapping FB2 tags to Markdown syntax, and can handle nested tags. The developer was motivated to build the tool due to the lack of suitable existing solutions, which were often bloated, required installation, or contained ads. The converter is fast, private, and correct, making it ideal for feeding text into large language models. The project demonstrates the importance of data transformation scripts in the era of AI and highlights the benefits of building custom tools to meet specific needs, rather than relying on existing solutions.