Supported formats
Docling can parse various documents formats into a unified representation (Docling Document), which it can export to different formats too — check out Architecture for more details.
Below you can find a listing of all supported input and output formats.
Supported input formats
| Format | Description |
|---|---|
| DOCX, XLSX, PPTX | Default formats in MS Office 2007+, based on Office Open XML |
| DOC, XLS, PPT | Legacy binary Office formats (97–2004); requires LibreOffice |
| RTF | Rich Text Format; requires LibreOffice |
| ODT, ODS, ODP | OpenDocument Format for text documents, spreadsheets, and presentations |
| EPUB | Electronic Publication format for e-books |
| Pages | Apple Pages documents (.pages), both container generations (Pages 5+ and iWork '09); requires the format-iwork extra |
| Numbers | Apple Numbers spreadsheets (.numbers), both container generations (Numbers 3+ and iWork '09); requires the format-iwork extra |
| Keynote | Apple Keynote presentations (.key), every container generation (Keynote 6+, the flattened package Keynote 2018+ writes, and iWork '09); requires the format-iwork extra |
| Markdown | |
| AsciiDoc | Human-readable, plain-text markup language for structured technical content |
| LaTeX | Scientific document preparation system |
| HTML, XHTML | |
| MHTML, MHT | MIME HTML archives |
| CSV | |
| PNG, JPEG, TIFF, BMP, WEBP | Image formats |
| WAV, MP3, M4A, AAC, OGG, FLAC | Audio formats (requires asr extra — see Processing audio and video) |
| MP4, AVI, MOV | Video formats — audio track is extracted and transcribed (requires asr extra and ffmpeg) |
| WebVTT | Web Video Text Tracks format for displaying timed text |
| BoxNote | Box Notes collaborative note format |
MIME (.eml) and Outlook (.msg) email messages; attachment names can optionally be listed via EmailBackendOptions |
|
| AFP | IBM Advanced Function Presentation / MO:DCA |
PowerPoint equations stored as Office Math (OMML) in text shapes are converted to LaTeX formula items. Equations mixed with text or list content retain their order.
For LaTeX input, front matter such as authors, affiliations, institutes, email
addresses, and dates is stored as text items. Inline formulas remain formula
children. The metadata, authors, and affiliations groups record source roles.
\thanks and \altaffiliation use footnote items in these groups. Retained
source commands use the furniture layer and do not appear in default Markdown.
Schema-specific support:
| Format | Description |
|---|---|
| DocLang XML | XML format following DocLang; supported extensions: .dclg, .dclg.xml, and generic .xml with a <doclang> root element |
| DocLang archive | Zipped DocLang bundle including page images; supported extension: .dclx |
| USPTO XML | XML format followed by USPTO patents |
| JATS XML | XML format followed by JATS articles |
| XBRL XML | XML format for business and financial reporting following XBRL standard |
| EBCDIC | Mainframe fixed-width data files; needs the COBOL record layout passed through EbcdicBackendOptions; supported extensions: .ebc, .ebcdic |
| Docling JSON | JSON-serialized Docling Document; images referencing local files are ignored unless enabled, see Docling JSON input |
Supported output formats
| Format | Description |
|---|---|
| HTML | Both image embedding and referencing are supported |
| Markdown | |
| JSON | Lossless serialization of Docling Document |
| DocLang XML | XML serialization following DocLang; CLI output format: doclang |
| Text | Plain text, i.e. without Markdown markers |
| Doctags | Markup format for efficiently representing the full content and layout characteristics of a document |
| WebVTT | Web Video Text Tracks format for displaying timed text |
| DocLang archive | Zipped DocLang bundle including page images; CLI output format: dclx |
| Chunks (JSONL) | Chunked document output for RAG pipelines; configurable via --chunks-type, --chunks-max-tokens, --chunks-tokenizer |
| LaTeX | Standalone .tex document; images are emitted as placeholders |