# Document conversion with Docling

# What's new

Docling runs as a managed API on the PHOENIQS AI Model as a Service platform. The service uses the vision-language model inference-qwen3-vl-235b to interpret each page visually, which supports high-quality conversion of documents with complex layouts.

# Why we built it

Enterprise files often contain tables, scanned pages or multi-column layouts that language models cannot use directly. Extracting clean, structured text from them usually means running your own parsing and OCR stack. We added a conversion service to the AI Model as a Service platform so that preparing documents and inferencing on them happen in the same place, under the same key and the same infrastructure.

# How it works

Docling is an open-source document conversion toolkit. It converts documents from their native formats into structured, machine-readable output such as Markdown, plain text, JSON and HTML. Unlike plain text extractors, it interprets document structure, including tables, reading order, headings and figures, and carries that structure through to the output.

Supported input formats: PDF, Word, PowerPoint, spreadsheets (Excel, CSV, ODS), OpenDocument files, HTML, Markdown, AsciiDoc, LaTeX, images (PNG, JPEG), XML, JSON Docling, DCLX, EPUB, email, Box Note, audio and VTT.

Supported output formats: Markdown, plain text, JSON, HTML, HTML split by page, YAML, DocTags, DocLang, DCLX, chunks and VTT.

# Why use it

No infrastructure to manage. Upload a file and receive structured output in seconds. Nothing to deploy, scale or maintain on your side.
No new credentials. Use the same API key you already use to inference PHOENIQS models.
Ready for downstream use. Output can be fed directly into RAG pipelines, search indexing or downstream LLMs.

# Pricing

There is no separate charge for document conversion. Usage is billed through your existing API key, based on the tokens consumed by inference-qwen3-vl-235b.

# How to use it

# Endpoint

POST https://maas.phoeniqs.com/v1/convert/file

# Example request

curl --location 'https://maas.phoeniqs.com/v1/convert/file'
    --header 'Authorization: Bearer <API_Key>'
    --form 'file=@"/path/to/document.pdf"'
    --form 'to_formats="text"'

# Required fields

Field	Description
`file`	The document to convert. Accepts PDF (.pdf), Word (.docx), PowerPoint (.pptx), spreadsheets (.xlsx, .csv, .ods), OpenDocument files (.odt, .odp), HTML, Markdown, AsciiDoc, LaTeX, common image formats such as PNG and JPEG, XML, JSON Docling, DCLX, EPUB, email, Box Note, audio and VTT.
`to_formats`	The output format. Accepts `md`, `text`, `json`, `html`, `html_split_page`, `yaml`, `doctags`, `doclang`, `dclx`, `chunks` and `vtt`.

# Optional fields

Field	Description
`page_range`	Converts selected pages only, for example: `--form 'page_range=1' --form 'page_range=10'`