Show HN: Was pissed about Google Docs, So I made an Text Editor myself (sourcepilot.co)

A developer frustrated with Google Docs built a minimalist native desktop text editor that emphasizes focused writing plus optional, locally-run AI assistance. The app offers a clean interface and instant, native performance for basic editing without any login; AI-powered suggestions and writing help are available only if users download a local model, keeping all inference and data on the user’s machine. For the AI/ML community this is notable because it pushes the "local inference" workflow: instead of relying on cloud APIs, the editor offloads LLM inference to user devices, improving privacy, ownership, and latency. That design highlights practical trade-offs — you get data control and lower round-trip delays, but you must manage model downloads, storage, and compute (CPU/GPU) constraints. The project underscores ongoing demand for compact, quantized models and optimized runtimes (e.g., GGML/ONNX-style toolchains) that make capable LLMs viable on desktops. It’s a small but clear example of how consumer tooling can shift toward on-device AI, encouraging more lightweight models, better local inference tooling, and editor integrations that prioritize privacy and responsiveness.