Blogg

AnyMD-blogg

Teknikk, guider og benchmarker om dokument-til-Markdown-konvertering, RAG-pipelines og raske Rust-API-er.












Veiledninger 26. august 2026 · 8 min lesing

Semantisk chunking fra Markdown for RAG

Markdown bevarer dokumentstruktur — overskrifter, lister, tabeller — noe som gjør det til ideell inngang for semantiske chunking-strategier. Sammenlign overskriftsbasert, token-basert og rekursiv chunking på AnyMD-utdata.

Les mer →


Guides 28 Aug 2026 · 7 min read

Token-aware chunking for LLM context windows

Split Markdown documents into chunks that fit exactly inside your model's context window — with overlap, heading boundaries, and metadata injection. Works with OpenAI, Claude, and Llama tokenizers.

Les mer →

Guides 31 Aug 2026 · 8 min read

Heading-based hierarchical chunking

Use Markdown heading structure (#, ##, ###) to create a hierarchy of chunks — parent sections contain their children. The best chunking strategy for document Q&A and summarization.

Les mer →