No more manual documents

Most business documents are assembled from information a company already has. An invoice, a standard contract, a monthly report: the names, numbers and dates are in a system, and the wording and layout barely change. Yet someone still exports the data, fills a Word template, fixes the formatting and emails the result.

I think that work should be fully automated.

If a document is built from structured data and a template, the system should generate it. The same goes for corrections. If a name is wrong, fix the data. If a clause needs changing, update the template. Then generate the document again.

This also makes a paperless process easier. Approval, electronic signing and delivery can happen digitally, without printing and scanning the document along the way.

I built docgen around this approach. It is a multi-tenant SaaS that turns JSON into PDF documents.

A template is created in a visual editor and published. Then any application can send data to the API and receive the finished PDF. For a certificate, that might be the course, the issuer, the date and the recipients. The template handles where they appear and how the document looks.

AI agents, working in parallel, wrote the specification, code, tests and infrastructure.

Try it

It is a demo, so please don't send real personal data to it.

  1. Sign up at docgen.vikonomov.com and open your project.
  2. Design a template in the editor, or publish a starter template. Test it in the Playground with sample data.
  3. Create an API client in the Test environment, keep the render:create permission and copy the key from the one-time dialog.
  4. Send a render request:
curl --request POST 'https://docgen.vikonomov.com/v1/renders' \
  --header 'Authorization: Bearer dg_test_REPLACE_WITH_YOUR_KEY' \
  --header 'Content-Type: application/json' \
  --data-raw '{
    "template": "certificate",
    "data": {
      "issuer": {
        "name": "Redkite Academy",
        "signatories": [{ "name": "John Doe", "title": "Office Manager" }]
      },
      "course": { "title": "Customer Service Excellence", "hours": 8 },
      "issueDate": "2026-10-13",
      "place": "Stonebury",
      "recipients": [
        {
          "name": "Jane Doe",
          "certificateId": "CERT-2026-0001",
          "completedOn": "2026-10-11",
          "verificationUrl": "https://verify.example.com/c/CERT-2026-0001"
        }
      ]
    }
  }' \
  --output certificate.pdf

The response body is the PDF itself.

Play with docgen, but treat it as a proof of concept. For real documents, build a similar approach into your own systems: pick one recurring document, specify its fields, wording and rules, and have AI agents implement it against that specification.

If a team still opens Word every time the next customer needs the same document, that is a good first one to automate.