[FEATURE] Prefer plain text over image when pasting from rich text sources (e.g. Word, Excel)
Preflight Checklist
- [x] I have searched existing requests and this feature hasn't been requested yet
- [x] This is a single feature request (not multiple features)
Problem Statement
When copying text from Microsoft Word or Excel (macOS) and pasting into the Claude Code chat panel in VS Code, the content is always pasted as an image. Currently, ⌘⇧V (paste as plain text) does not work in the chat panel either.
Proposed Solution
Implement clipboard format detection that prefers text/plain over image when both representations are available. Ensure ⌘⇧V works in the chat panel as a paste-as-plain-text fallback.
Alternatively, a setting like claudeCode.pastePreference: "text" | "image" | "auto" would let users control this behavior.
Alternative Solutions
Tried:
- ⌘⇧V (paste as plain text) – does not work in the chat panel, still pastes as image
- VS Code setting editor.pasteAs.preferredType: "text" – has no effect on the chat panel (it's a webview, not a standard editor)
- Pasting through an intermediary (e.g. TextEdit, browser address bar) – works, but defeats the purpose of a seamless workflow
Currently using:
There is no viable workaround. We either paste via an intermediary app or skip pasting altogether and ask Claude to fetch the file directly from SharePoint via MCP.
Priority
High - Significant impact on productivity
Feature Category
Interactive mode (TUI)
Use Case Example
Marketing and content teams frequently draft in Word/Google Docs, work with data from Excel, and paste excerpts into Claude Code for review, editing, or processing. The current behavior provides no way to paste as plain text – the chat input always interprets clipboard content from these sources as an image.
Additional Context
Environment
macOS (Darwin 24.6.0)
VS Code with Claude Code extension
Source: Microsoft Word (.docx)
<img width="458" height="137" alt="Image" src="https://github.com/user-attachments/assets/94bd4142-d6d7-4139-8686-2cbd48261a43" />
This issue has 6 comments on GitHub. Read the full discussion on GitHub ↗