> ## Documentation Index
> Fetch the complete documentation index at: https://docs.markpdf.tech/llms.txt
> Use this file to discover all available pages before exploring further.
# Supported formats
> What enters, how it is detected and how the ZIP works.
# Supported formats
## Entries
| Format | Extension | Notes |
| ---------- | --------- | ------------------------------------------- |
| Word | `.docx` | |
| Excel | `.xlsx` | |
| PowerPoint | `.pptx` | |
| CSV | `.csv` | Converts to Markdown table. |
| Text | `.txt` | It is returned as is, normalized. |
| ZIP | `.zip` | Container with a document supported inside. |
## Auto detection
With `input_format=auto` (default) the order is:
The extension of `filename` is looked at.
If the extension is not enough, the header `content-type` is used.
If nothing matches, it responds `415` asking you to set `input_format`.
If you know the format, set `input_format` explicitly. It is faster and avoids ambiguities.
## ZIP
When you send a ZIP, **the first supported document** it finds is extracted.
* Hidden folders and files (those starting with `.`) are ignored.
* If ZIP has more entries than allowed, `413` responds.
* If no member is a supported format, reply `415`.
* If the internal document exceeds the size limit, respond `413`.
The ZIP is for packaging **one** document, not batch processing. For multiple documents, make a request for each one.
## CSV to Markdown
The CSV is converted to a Markdown table. Cells are escaped (`|` and line breaks) and trimmed if they are very long, to keep the table readable.
## What is preserved and what is lost due to formatting
Converting to Markdown always means losing some visual fidelity — Markdown
It has no concept of a page, exact source or absolute position. This table
summarizes what survives reasonably well and what is flattened or discarded:
| Format | It is preserved | Is lost or flattened |
| ---------------------- | --------------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
| **Word (.docx)** | Headings (`Heading 1`-`Heading 2` -> `#`-`##`), bold/italics, numbered and bulleted lists, tables, links. | Comments and track changes, footnotes are flattened at the end or lost depending on the converter, headers/footers, section breaks. |
| **Excel (.xlsx)** | Each sheet is converted to a Markdown table; formulas are solved to their calculated values. | Formulas themselves (just the result), conditional formatting, embedded graphics, multiple sheets are concatenated sequentially (there are no "tabs" in Markdown). |
| **PowerPoint (.pptx)** | Text on each svalide in order, titles as headings, bulleted lists. | Visual design, position of text boxes, images and graphics (only the accompanying text), animations and transitions. |
| **CSV** | All rows and columns as Markdown table; Long cells are cropped for readability. | Data types (everything becomes text), numeric format/original date. |
| **TXT** | The full text, normalized (line breaks, spaces). | Not applicable — there is no structure to lose. |
| **ZIP** | The internal supported document is converted the same as if it had been uploaded loose. | Everything else within ZIP (only the first supported document found is processed). |
If your use case depends on preserving complex tables or multi-column layouts
Complete fidelity comparison by mode.
## Real use cases by format
* **PDF + `fast`**: bulk ingestion for a RAG search index where
Speed matters more than preserving every nuance of layout.
* **DOCX for legal contracts**: `clean=true` (default) is especially useful
here because contracts usually have repeated headers/footers in each
page that dirty the Markdown if they are not normalized.
* **XLSX of financial reports**: each sheet becomes a Markdown table
independent — useful for a LLM to reason about a sheet of "Q3
Revenue" without having to parse a binary file.
* **PPTX from sales decks**: extracting only the text from the svalides is
enough for an agent to summarize the content of a presentation without
need to process images.
## Troubleshooting
| Symptom | Cause | Solution |
| ------------------------------------------------------------- | ----------------------------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| `415` with a file that "should" work | The extension does not match the actual content, or the client's `content-type` is generic (`application/octet-stream`) | Set `input_format` explicitly instead of relying on `auto`. |
| Excel tables come out as plain text instead of Markdown table | The sheet has merged cells or a very irregular layout that the converter cannot map to clean rows/columns | Simplify the layout of the source sheet, or accept the output as text and post-process. |
| ZIP cannot find any vavalid documents | The files inside the ZIP are in nested folders further than expected, or they all start with `.` (hidden) | Verify that the supported document is in the root or a visible folder of the ZIP. |
| The CSV comes out with misaligned columns | CSV uses a non-comma delilimitr (`;`, ` `) without declaring it | Normalize the CSV to comma delilimitr before uploading it, or use explicit `input_format=csv` to force the correct parser if your CSV uses consistent quotes. |