[BUG] docs_get_document returns no content for Google Docs tables
Preflight Checklist
- [x] I have searched existing issues and this hasn't been reported yet
- [x] This is a single bug report (please file separate reports for different bugs)
- [x] I am using the latest version of Claude Code
What's Wrong?
Description
The docs_get_document tool returns paragraph elements only. Any content inside a Google Docs table is silently omitted -- no error, no placeholder, just a gap in the character indices.
## Actual Behavior
Table cells are entirely absent. The character index gaps in the response confirm the tables exist in the document but their content is not exposed. For example, a document with a table between
indices 10,062 and 10,616 returns nothing for that range.
## Impact
Documents that use tables for structured content (e.g., data mappings, reference tables, decision matrices) are effectively unreadable via this tool. There is no fallback — drive_export_file as
HTML returns a non-Google-Workspace file, which drive_get_file cannot read either.
## Environment
- API Platform: claude.ai web app with Claude Code CLI
- Tool access: G-Suite MCP connector configured via claude.ai
- Connector: G-Suite MCP (claude.ai built-in connector)
- Tool:
docs_get_document - Tested on: Google Docs with single and multi-column tables
What Should Happen?
## Expected Behavior
Table cell content should be included in the response, either inline with paragraphs (preserving reading order) or as a separate tables field in the response.
Error Messages/Logs
Steps to Reproduce
## Steps to Reproduce
- Create or open a Google Doc that contains one or more tables with text content.
- Call
docs_get_documentwith the document's ID. - Inspect the returned
paragraphsarray.
Claude Model
Sonnet (default)
Is this a regression?
I don't know
Last Working Version
_No response_
Claude Code Version
2.1.77 (Claude Code)
Platform
Other
Operating System
macOS
Terminal/Shell
iTerm2
Additional Information
_No response_
This issue has 2 comments on GitHub. Read the full discussion on GitHub ↗