built Web2JSONL to solve a problem I faced repeatedly: converting disparate sources (websites, documents, images) into clean, training-ready JSONL datasets for AI models. Whether you're fine-tuning an LLM, building a custom model, or preprocessing data for AGI research, Web2JSONL eliminates the tedious data wrangling. You can paste text directly, scrape web pages with a single URL, or upload multi-format files with OCR support. Everything flows through our staging buffer, where you can organize and prepare data for training. The tool supports three output formats—pretrain, instruct, and chat—optimized for different training paradigms. It's built with developers in mind: minimal UI, maximum functionality, zero bloat. v1.0.5 is a lean, focused release. We're already planning: 🔮 Batch processing for multiple URLs 🔮 Advanced data filtering & cleaning 🔮 Dataset versioning & export history 🔮 API for programmatic access If you're working on LLM training datasets or AGI projects, give it a try. I'd love to hear how you use it!"

Web2JSONL

Web2JSONL – Convert Websites to JSONL for AI Model Training

Web2JSONL – Convert Websites to JSONL for AI Model Training