Show HN: Writekin – fine-tune a local LLM on your own writing, on your Mac (github.com)

Writekin has introduced a novel tool that allows users to fine-tune a local language model based on their own writing style directly on Mac devices. By ingesting text from emails, messages, and other documents, Writekin curates this data into a training corpus for a local QLoRA model, enabling the generation of text that mirrors the user's unique voice. Notably, all processing occurs locally, ensuring user privacy—no data leaves the user’s machine, and there’s minimal network traffic related only to necessary model downloads. This development is significant for the AI/ML community as it democratizes access to personalized AI tools, allowing individuals to harness the power of advanced language models without compromising their data. Designed for Apple Silicon Macs and requiring a minimum of 16 GB of unified memory, Writekin caters to personal use while also providing a transparent source-available codebase. It supports seamless integration of user content, promotes a non-commercial license model, and adheres to privacy-centric principles, setting a precedent for user-focused, locally operated AI solutions.