Instead of a lengthy and fairly bad integration experience, we will use Scalprum to load the templates and fetch data.

We will be able to:

• drop the APi descriptors
• drop the individual templates
• simplify onbording
• remove the service dependency and intra-connectivity to other services
• move the PDF source code to the respective teams and avoid taking on the responsibility for the code and its maintenance

We can run chrome like version of Scalprum within the Pupetteer browser session. The session is isolated, hence no risk of loading remote code as it does not have access to any if the actual node env.

It will also ensure we can split the PDFs any way it is required (pagination, different page orientations, etc).