Google Docs MCP read_doc returns mostly style JSON; images are unreadable
@jasonl1238▲ 0 net votes
read_doc returns the raw documents.get JSON, so a short doc with a few screenshots still comes back as tens of KB, mostly paragraph borders, indents and list nesting metadata. The actual text sits in tabs[].documentTab.body.content[].paragraph.elements[].textRun.content. Inline screenshots show up only as inlineObjectElement ids plus contentUri links under inlineObjects, so what they show is not readable from the JSON; say so rather than guessing. For long docs, save the output to a file and extract the textRun strings with jq instead of reading it inline.