System One
Send a state and typed questions to Solar Jev and get back a choice, a score, or a yes/no answer, each with the probabilities the model assigned. Covers the request parameters, the answer objects, and an example request and response.
Asynchronous API
Process documents with up to 2,000 pages asynchronously. Submit requests and retrieve results in batches.
AuthorizationIn: header
multipart/form-dataRequiredmodelA string value indicating the model used for inference.
We recommend using the alias document-parse, which always points to the latest stable model.
You may also specify a particular model version, but please note that models are frequently updated, and services tied to specific versions may be affected when those versions are deprecated.
Available snapshot models are listed in the capabilities document.
Use document-parse-nightly to test the latest features. This model may be updated without notice.
documentThe document file to be processed. Supported file formats are listed here.
modestringA string value specifying the parsing mode. Note that enhanced and auto modes are available from document-parse-260128 and later versions.
standard for text-focused documents with simple tables.enhanced for documents containing complex tables, images, charts, and other advanced visual elements.auto to automatically classify pages as either standard or enhanced and process them accordingly."standard"Value in: "standard" | "enhanced" | "auto"chart_recognitionA boolean value indicating whether to use chart recognition. If true, charts are converted to tables.
When mode is set to enhanced, chart recognition is always enabled regardless of this parameter.
trueValue in: true | falsemerge_multipage_tablesA boolean value indicating whether to merge tables that are split across multiple pages into a single table. When set to true, tables spanning multiple pages will be combined into one cohesive table structure.
falseValue in: true | falseocrstringA string value indicating whether to perform OCR inference on the document before layout detection. When set to auto, OCR is applied only to image documents; for non-image files (e.g., PDF, DOCX), layout and text detection will work only if the document is digital-born. When set to force, the input file is converted to images, and OCR is always performed before layout detection.
"auto"Value in: "auto" | "force"output_formatsstringA string value indicating in which each layout element output is formatted. Possible values are text, html, and markdown.
"['html']"coordinatesbooleanA boolean value indicating whether to return coordinates of bounding boxes of each layout element. The default value is true.
trueValue in: true | falsebase64_encodingstringA string value indicating which layout category should be provided as a base64 encoded string. All category names can be found in the Layout Categories and HTML Tags section. This feature is useful when users want to crop the layout elements from the original document image and store or use them for their own purposes. For example, users can extract the base64 encoded images of all tables in the input document by setting base64_encoding=["table"]. In the response, the elements[].base64_encoding field will contain the base64 string of the corresponding element image. Note that this may increase the size of the response.
[]wordsA boolean value indicating whether to include word-level OCR results in the response. When set to true, each recognized word and its bounding-box coordinates are returned in addition to the layout elements. Note that this may increase the size of the response.
falseValue in: true | false