Feature Request: Support image paste/drag-drop input in VSCode extension chat
Resolved 💬 2 comments Opened Jun 16, 2026 by cangs007 Closed Jun 20, 2026
Summary
Add the ability to paste images (from clipboard) or drag-drop image files (PNG/JPG) into the Claude Code VSCode extension chat input, so Claude can read and process image content directly — similar to the web version and other AI coding assistants.
Motivation
Currently, the VSCode extension only accepts text input. When users need Claude to analyze screenshots, UI mockups, diagrams, error dialogs, or other visual content, they must:
- Save the image to a file
- Manually type the file path into the chat
- Ask Claude to use its file reading tool to view the image
This is a significant UX friction compared to:
- The Claude.ai web version, which supports direct image upload
- Competing tools (GitHub Copilot, Cursor, etc.) that support in-chat image input
Use Cases
- Bug analysis: Paste a screenshot of an error dialog or crash message
- UI development: Show a mockup/design and ask Claude to implement it
- Visual debugging: Share a rendering issue or layout problem
- Diagram interpretation: Architecture diagrams, flowcharts, wireframes
- Code in screenshots: Sometimes code is shared as images in review contexts
Suggested Implementation
The most straightforward approach would be one or both of:
- Clipboard paste (Ctrl+V): When the user pastes an image from clipboard into the chat input, encode it and attach it to the conversation context
- Drag-and-drop: Allow dragging image files (
.png,.jpg,.gif,.webp) from the file explorer into the chat input area
Ideally the image is converted to a format Claude's Read tool can process, or sent as a base64-encoded image in the message.
Additional Context
- Platform: Windows 11 / VSCode
- Extension version: latest
- This feature already exists in the web version of Claude, so it's about porting it to the VSCode extension
Related
- The Read tool already supports reading/displaying PNG and JPG images via file path — this request is about making the input flow seamless for users.
This issue has 2 comments on GitHub. Read the full discussion on GitHub ↗