Nano PDF Released: A CLI Tool for Editing PDFs Using Gemini 3 Pro Image

Nano PDF works by converting the target page for editing into an image and sending it to the Gemini 3 Pro Image API along with a prompt. The resulting edited image is then synthesized back into the original PDF for output.

Users can perform edits by specifying the page number and instructions, for example: nano-pdf edit deck.pdf 5 "Update the revenue chart to show Q3 at $2.5M".

Additionally, the tool supports adding new slides. It generates new content that matches the style of existing slides and inserts it into the document.

Parallel editing of multiple pages is possible, and Google Search is enabled by default, allowing the model to reference the latest data for its edits.

Because the text layer is preserved, copy-and-paste and search functions continue to operate normally.

The GitHub repository can be found here.