Supported File Types

The Read File component handles a range of formats. Which formats are accepted in any given configuration depends on the chosen extraction backend and, for Upload Targets, the Allowed Upload Extensions field.

Standard File Support

These are inherited from the upstream Langflow Read File component (the docling backend).

KARLI Extraction Support

When Extraction Backend is set to karli, the supported formats depend on the selected model:

Model	Accepts
karli/default-data-extraction	Any supported file (the service picks an extractor).
karli/data-extraction-moe-latest	Any supported file (MoE router picks the optimal extractor per file type and per page for PDFs).
docling-project/docling, datalab-to/marker, opendatalab/MinerU, karli/multimodal-data-extraction	Document formats (PDF and similar).
openai/whisper-large-v3	Audio formats.

The component validates the file against the chosen model's accepted type before uploading.

The MoE model supports a broad range of formats including PDF, Word (doc/docx), presentations (ppt/pptx), spreadsheets (xls/xlsx/csv), HTML, images (png/jpg/gif/bmp/tiff/webp), audio (aac/mpeg/wav/webm/mp3/mp4), email (eml/msg/pst), and plain text. See Models → Data Extraction for the full routing table.