View only. Sign in to run this demo. You can also run it with your own model key, without an account.
Document → Structured Data
Turn a document or an image into structured data: it is parsed to text, and the fields you describe are extracted from it
How to use Document → Structured Data
- 1Add a documentUpload a PDF, Word file or image, or start from a ready-made document with the parser and task that suit it.
- 2Choose the parserThe parser turns the file into text. A text PDF needs a fast one; a scan or a photo needs a vision model.
- 3Say what to extractDescribe the fields in plain words, including lists. The task decides the shape of the result.
- 4Check the resultRead the data, then the text it came from, and the schema that was made for it.
- Three steps run in turn: the document is parsed to text, a schema is made from your task, and the data is extracted from the text. If the parser misses something, the data does too, so read the parsed text.
- Nothing you enter is saved. The examples come with a ready-made schema. A schema made for your own task, or for an edited example, lives in this session only and is used again while the task stays the same.
Start from a ready-made document
1. The document
A PDF, Word file or image, up to 20 MB.
2. The parser
How the file is turned into text. The ready-made documents set the one that suits them.
This parser is good for fast, high-quality parsing of PDFs with headings and tables. It uses the GPU-accelerated Docling service hosted by Haaga-Helia and returns markdown with metadata.
3. What to extract
Describe the fields in plain words. Ask for a list, and you get one row for each item.
This task has a ready-made schema, so no time goes on making one. Change the text and a new schema is generated.
A PDF with the extracted data, to download after the run.
Add a document first.
No structured data yet
Add a document and click Extract data. The data appears here, with the text it came from.
