Great concept
Poppy AI has a lot of things I genuinely like. The visual board approach is very useful, especially when working with different sources and trying to organize research, ideas, documents, videos, and content in one place. The overall concept is excellent for content creation, research, marketing, educational material, and building workflows around source material.

One area, however, seriously weakens the experience: the way credits are consumed when working with attached documents.

When a PDF remains connected to a chat, its content is processed again with every new AI message. THIS MEANS THAT A LONG DOCUMENT CAN REPEATEDLY CONTRIBUTE TO CREDIT CONSUMPTION EVEN WHEN THE SOURCE ITSELF HAS NOT CHANGED.

For users who work with large PDFs, manuals, educational material, research papers, transcripts, or other long-form resources, this can make credits disappear surprisingly quickly. A workflow that naturally requires several follow-up questions can become much more expensive than expected simply because the same source content is being processed repeatedly.

This is not just a small pricing detail. It directly affects how freely and effectively the platform can be used for serious research and content workflows.

Poppy would be significantly stronger IF UPLOADED DOCUMENTS COULD BE INDEXED ONCE and then referenced more efficiently in subsequent questions, RATHER THAN REPEATEDLY CONSUMING CREDITS for the SAME UNCHANGED MATERIAL.

The product itself has a very strong concept and a lot of potential, and the visual workflow is one of its biggest strengths. But the current document-related credit consumption is an important weakness that should be improved, because it can make otherwise normal usage unnecessarily expensive.

. This is the ONLY but very serious reason what i give 3 stars. When you cure this please let me know and immediately I will raise my review to 5 stars.