Why convert PPTX before using ChatGPT?
PowerPoint files often mix visible text, speaker notes, tables, chart data, images, SmartArt, and native shapes. A structured package lets you inspect what was extracted before sharing it with an LLM.
Recommended flow
Convert the presentation into an LLM package, then start with content.md for most ChatGPT tasks. Keep content.compact.md for smaller context windows, and use manifest.json to understand which files are available.
Which output to use
Markdown is best for direct reading and summarization. HTML is useful for human review. JSON is better for agents, RAG pipelines, validation, and automated workflows that need structured slide data.
Images and speaker notes
Content images are stored under assets/ and referenced from Markdown and HTML. Speaker notes are extracted into JSON, Markdown, compact Markdown, and HTML when the PPTX includes notesSlides data.
Quality checks before sharing
Use the Preview panel and Details workspace to inspect slide summaries, generated files, conversion warnings, package contents, and whether tables, charts, images, notes, or shape text were extracted.
Privacy note
The PPTX converter runs in the browser by default, so presentation files do not need to be uploaded to a server for the ready workflow.
FAQ
Does the Markdown include every image?
Markdown references content images by relative path. Repeated logos and decorative images may be classified separately so they do not overwhelm the LLM prompt.
Should I use Markdown or JSON?
Use Markdown for ordinary reading. Use JSON when an agent or script needs stable fields for slides, tables, charts, notes, images, and warnings.
Is the ZIP different on Markdown, HTML, and JSON pages?
No. PPTX pages generate the same complete ZIP package. Each page only chooses a different default preview tab.