Skip to main content

Parsing modes

LlamaParse support different parsing modes adapting to your use case. The three main modes are:

  • fast mode, that prioritize speed
  • balanced mode that balance cost and accuracy
  • premium mode that prioritize accuracy.

If you need more control over your parsing, you can also use our Advanced parsing mode

Fast Mode​

Fast mode bypass reconstruction on your document, greatly accelerating parsing.

The Markdown output will not be generated.

Fast mode is billed .1 cent per page, but at least .3 cent per document.

To use the fast mode, set fast_mode to True.

In Python:
parser = LlamaParse(
  fast_mode=True
)
Using the API:
curl -X 'POST' \
  'https://api.cloud.llamaindex.ai/api/parsing/upload'  \
  -H 'accept: application/json' \
  -H 'Content-Type: multipart/form-data' \
  -H "Authorization: Bearer $LLAMA_CLOUD_API_KEY" \
  --form 'fast_mode="true"' \
  -F 'file=@/path/to/your/file.pdf;type=application/pdf'

Balanced Mode​

Balanced mode (previously called 'accurate') try to balance cost and accuracy. It is a good compromise for most documents.

Balanced mode is billed .3 cent per page (3usd / 1000 pages)

To use the balanced mode, nothing is needed in the API as it is our default mode.

Premium Mode​

Premium mode leverages state-of-the-art multimodal models and heuristic text parsing techniques to extract text from the most complex documents with our best accuracy possible.

Fast mode is billed 4.5c per page.

To use the premium mode, set premium_mode to True.

In Python:
parser = LlamaParse(
  premium_mode=True
)
Using the API:
curl -X 'POST' \
  'https://api.cloud.llamaindex.ai/api/parsing/upload'  \
  -H 'accept: application/json' \
  -H 'Content-Type: multipart/form-data' \
  -H "Authorization: Bearer $LLAMA_CLOUD_API_KEY" \
  --form 'premium_mode="true"' \
  -F 'file=@/path/to/your/file.pdf;type=application/pdf'