[FEATURE] Support image/screenshot paste in Claude Code terminal
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 working with Claude Code for data-heavy or visual workflows, I often need to provide screenshots as context — such as brokerage transaction history captures, chart images, UI mockups, or error dialogs. However, Claude Code currently doesn't support pasting or dropping images directly into the terminal interface.
When I'm debugging or building features based on visual references, it is difficult to convey the exact context without image support. Describing a screenshot in text is time-consuming and often loses important details.
The current workflow requires me to manually save the image as a file, note the file path, and ask Claude Code to read it via a separate command — which significantly breaks the flow compared to simply pasting a screenshot.
Proposed Solution
Add support for image input in Claude Code via:
- Ctrl+V paste of clipboard images directly into the prompt
- Drag and drop of image files onto the terminal window
- Inline image preview (thumbnail) before submission so the user can confirm the correct image was attached
This would align Claude Code's image handling with the capability already available in claude.ai web/desktop chat, where images can be pasted or dragged in seamlessly.
Alternative Solutions
I am building a personal portfolio management web app using Claude Code. My workflow involves referencing brokerage screenshots (transaction history, account balance captures) to guide feature implementation. Being able to paste these captures directly into Claude Code — just like in the web chat — would dramatically improve productivity and reduce context loss.
Claude's vision capability is already strong in the web interface. Bringing that same experience to Claude Code would make it significantly more powerful for users working with visual data, UI work, or any screenshot-heavy debugging workflow.
Priority
High - Significant impact on productivity
Feature Category
Performance and speed
Use Case Example
_No response_
Additional Context
_No response_
Showing cached comments. Read the full discussion on GitHub ↗
15 Comments
Found 3 possible duplicate issues:
This issue will be automatically closed as a duplicate in 3 days.
🤖 Generated with Claude Code
+1 on this feature request.
+1
+1
i found that it can be done , at list in windows with Ctlr + Alt +V , otherwise you can search the "Terminal Image Paste" vscode extension , granted you started claude inside the VS code terminal
Windows 11 solution: https://github.com/anthropics/claude-code/issues/12644#issuecomment-4187771042
For macOS users: clipaste solves this today. Background daemon that auto-saves screenshot clipboard images to temp files and registers the file URL, so both Cmd+V and Ctrl+V work in any terminal.
Works with Ghostty, Alacritty, iTerm2, Terminal.app + Claude Code, Codex CLI, Cursor.
+1
Yes!!
Dear anthropic, HOW LONG!!!!?
How long will I have such a mess on my Desktop just to send you the image via drag and drop instead of damn Ctrl+V!!!?
<img width="780" height="773" alt="Image" src="https://github.com/user-attachments/assets/7de2a073-0045-47e8-a274-eb14b1a50a40" />
1+
+1
I installed the Playwright MCP server plugin and I have claude take the screenshots for me.
Paste support is the obvious half. The half that bites after it lands: on a dense screen the model still has to work out which element you mean, and "the second field in the third card" reads as nothing to it. The inline thumbnail in your third bullet confirms it for the human, but the model would get further if the paste carried which element you marked plus its position and text, not just the flat image. Otherwise you paste, it edits the wrong thing, you re-prompt.
If you're on macOS, I built cc-preview which might help in the meantime.
Once you've pasted images (they show up as
[Image #N]tags), cc-preview lets you browse and preview them from past sessions using a native Hammerspoon UI — so you can always refer back to what you pasted.Then press Cmd+Shift+I to open it.
Nice, cc-preview looks handy for the refer-back case, seeing what you already pasted. That's a different part of the flow from the one I was poking at, though. cc-preview helps after the paste. The gap I keep hitting is before it, the paste carrying which element you marked plus its position and text, so the model doesn't re-guess on a dense screen. I built slimsnap.ai for that half (Mac, and the handoff format is open). Two ends of the same flow really, mark-and-send upstream, browse-what-you-sent downstream.