Use case
Knowledge bases (RAG)
Clean, chunked text for retrieval. From pages or documents.
The problem.
Retrieval quality starts with the text you index. Navigation poisons an index. So do PDFs with broken layout. The damage is done before the first query.
How Webfold helps.
Step 1
Convert
Web pages, PDFs, Word and Excel files all become Markdown.
webfold convert report.docxStep 2
Chunk
Split into retrieval-sized pieces with stable boundaries.
webfold chunk report.pdf --size 4000 --jsonlStep 3
Write to your store
Send NDJSON records to JSONL, Parquet or Iceberg.
webfold sink parquet < records.ndjsonProducts used.
Worth knowing.
Note
- Always spot-check extraction on your own content. Measured quality is on the Benchmarks pages.
Try it on your own data.
Run the command-line tool today. Hosted API access is coming.