Projects

Whole documents. Structured possibilities.

Make a project for each topic, client, or piece of work and upload its documents. A vision-language model (VLM) reads every page and turns it into structured JSON — use it as context in AI chat, or download it and use it anywhere.

Open the file. Discover what is inside.

Projects / OperationsPrepared example

operations-review.pdf · page 1

Quarterly operations

Gross margin increased during Q3.

Metric	Q2	Q3
Gross margin	41.0%	47.2%

Review the surrounding assumptions.

{
  "document": "operations-review.pdf",
  "pages": [
    {
      "actual_page_number": 1,
      "page_summary": {
        "medium_summary": "Quarterly operating metrics and review notes.",
        "keywords": [
          "operations",
          "margin",
          "quarterly review"
        ]
      },
      "extracted_text": {
        "heading": "Quarterly operations",
        "paragraph": "Gross margin increased during Q3.",
        "table": {
          "columns": [
            "Metric",
            "Q2",
            "Q3"
          ],
          "rows": [
            [
              "Gross margin",
              "41.0%",
              "47.2%"
            ]
          ]
        }
      }
    }
  ]
}

One document. Three useful next steps.Explore contentChat with itDownload JSON

Prepared example of page-level content. Extraction quality depends on the source; review incomplete or uncertain results.

From stored files to usable knowledge.

Content with its structure intact.

Explore extracted page content, summaries, and tables alongside the original file. Page references keep the result connected to where it came from.

A library organized around your work.

Group files into projects and folders, find content within extracted pages, and keep related material together.

Ready for the next question.

Start a chat from a ready file or scope a conversation to a folder or project. Use the processed documents as context for your questions.

Structured JSON you can take with you.

Download a ready document’s extracted JSON and feed it to your own tools, scripts, or other AI apps. Inspect processing issues and retry affected pages where the app offers that action.

Take the structure with you.

Document JSON, ready for your next step.

In Projects, download the extracted JSON of a ready document. The sample includes page content and a summary.

Download sample JSON

Put your documents to work.

Group documents by project, let the VLM structure every page, then ask in chat or take the JSON with you.

Try it out