Parsing creates machine-ready text from PDFs, images, or HTML snippets.

When to use parsing

• Feed OCR’d text into LLM prompts.
• Generate embeddings for semantic search.
• Pre-process files before custom NLP pipelines.

Studio workflow

1. Upload a document at https://app.algorythmos.fr.
2. Select Parsing and choose the output format: plain text or structured Markdown.
3. Enable Preserve layout if downstream systems rely on page markers.

API example

curl -X POST https://api.algorythmos.fr/parse/file \
 -H "x-api-key: $ALG_KEY" \
 -F "file=@/path/to/report.pdf" \
 -F "format=markdown"

The response returns a parser_run_id. Poll GET /parse/runs/{parser_run_id} until status becomes succeeded, then download the parsed content from the output_uri.