Korean Plain Text Dataset Needed for AI Training

We are looking for large-scale, original Korean text datasets for AI and LLM training. We welcome both ready-made datasets and vendors who can source and process eligible Korean text data. 📌 ACCEPTED CONTENT CATEGORIES 📚 Literature ▪️ Web novels ▪️ Fiction and fan fiction ▪️ Traditional stories ▪️ Literary works 💬 Forums ▪️ Complete discussion threads ▪️ Original posts and associated replies ▪️ Content reflecting natural Korean language and local usage ✍️ Blogs ▪️ High-quality personal blog posts ▪️ Long-form articles ▪️ Personal essays and opinion pieces ✅ STRICT DATA REQUIREMENTS ▪️ Content must have been originally written in Korean. ▪️ Machine-translated, manually translated, AI-generated, or synthetic content is not accepted. ▪️ Delivery format must be plain text in UTF-8 encoding. ▪️ PDFs, scanned documents, screenshots, images, and OCR-extracted files are not accepted. ▪️ Data must be clean, readable, and directly usable. ▪️ Duplicate and near-duplicate content should be removed or clearly identified. ▪️ The provider must be able to explain the source and collection method. ▪️ Applicants must confirm that the proposed data can legally be collected and provided for the intended use. ▪️ Source URLs or provenance records should be available upon request. 📩 HOW TO APPLY Please provide the following information: 1️⃣ Content categories you can provide: literature, forums, blogs, or a combination. 2️⃣ Whether the data is ready-made or needs to be newly sourced. 3️⃣ Estimated raw data volume in GB / Estimated usable token volume after cleaning and deduplication. 4️⃣ Proposed rate per GB or per 1 million usable tokens / Estimated lead time. 5️⃣ A representative plain-text sample of approximately 1–5MB, or a shareable Google Drive link. ⚠️ IMPORTANT Please do not submit machine-translated, AI-generated, OCR-extracted, duplicated, or unverifiable datasets. We are looking for vendors who can support large-volume delivery. Partial-volume proposals are also welcome.