Screenshots, sized for your AI coding assistant
Show Claude Code, Codex CLI, or Gemini CLI what you mean. Maxisnap captures a region or a screen recording on Windows and exports a one-click AI-optimized copy — right-sized so the model reads it clearly without burning tokens on 4K pixels it will only downscale.
Windows 10 (1903+) / 11 · macOS 13+ · free tier
Coding with an AI assistant in the terminal is a visual job more often than it looks. A stack trace in a dialog, a layout that renders wrong, a Figma frame you want built, a chart the model should read — the fastest way to explain any of them is a picture.
Maxisnap solves the picture, not the plumbing. It does not auto-drive the agent or paste for you — it gives you a clean, correctly-sized image (or a short recording) in one keystroke, so the paste, drag, or file-path step that follows is quick and reliable. Below is what each tool's own documentation says about adding an image, and where a purpose-built export earns its place.
Why shouldn't I just paste a full-resolution screenshot?
Because a full-resolution capture is usually more image than the job needs. Anthropic's vision guide, read in October 2026, says images above a model's size limits are downscaled before processing, and suggests downsampling images before sending them to control token costs when you do not need the extra fidelity. Each assistant sets its own image limits, so check the documentation of the one you use.
How do I give a screenshot to each CLI?
Each of the three documents its own ways to add an image, and Maxisnap's job is to hand you a ready-to-use file for whichever one you pick. This is what each tool's documentation said when we read it in October 2026:
- Anthropic's Claude Code documentation, read in October 2026, lists three ways to add an image: drag and drop it into the Claude Code window, paste it with Ctrl+V (Alt+V on Windows and WSL), or give Claude the image's file path.
- OpenAI's Codex CLI documentation, read in October 2026, says you can paste an image into the interactive composer or pass image files on the command line with --image.
- The Gemini CLI documentation, read in October 2026, says its read_file tool takes a file path and supports text, images, audio, and PDF.
Check each tool's own documentation for its current behaviour. Maxisnap only produces the image — it does not modify or automate the CLI.
What's the fastest way to hand a UI bug to Claude Code?
Capture the relevant area, remove sensitive content, and review the exported copy:
- Capture it. Press
Ctrl+Alt+5and drag a box around the error dialog, the broken layout, or the design frame. - Remove anything private. Keep API keys, tokens, and private customer data outside the capture or crop them out. If cropping is unsuitable, use a fully opaque redaction tool in another editor and inspect the exported file. Add an arrow or a numbered pin to point the agent at the relevant spot.
- Export the AI-optimized copy. One click produces a copy sized to a default max width of 1600 px (configurable in Settings). Save it, copy it, or drag it straight into the terminal — then paste, drop, or reference its path in Claude Code, Codex, or Gemini.
AI Mode reduces image dimensions before export, which can help with an assistant’s upload limits. Check the target tool’s current limits and inspect the image you will send. Keep credentials outside the capture or crop them out; blur alone does not guarantee that a secret cannot be inferred.
$ diff raw-screenshot ai-optimized
[RAW 4K] paste a full-monitor capture -- more pixels than the task needs [MAXISNAP] export AI-optimized copy ++ ~1600px wide, reads clean [RAW 4K] secrets on screen -- keys and tokens sent as-is [MAXISNAP] crop before export ++ crop secrets out and inspect the exported image [RAW 4K] image only on the clipboard -- no file to fall back on if a paste does nothing [MAXISNAP] saved file + drag-out ++ drop the file or reference its path
Do I need a dedicated tool, or can I just snip and paste?
You can absolutely snip and paste manually — for the occasional screenshot it's fine. A dedicated export earns its keep once you're doing it all day, care about token cost, and want to redact before sending. Here is what Maxisnap's AI Mode includes:
| What you want | Maxisnap AI Mode |
|---|---|
| Right-sized for tokens | 1600 px default, configurable |
| Runs on Windows | Yes |
| Blur a selected region | Blur tool (B) |
| Share a screen recording with AI | AI-optimized MP4 |
| Also a full screenshot + recorder | Both, one app |
| Price | Free tier |
For Maxisnap, this table describes the Windows app. A Mac version is also available, for macOS 13 or later.
Can I share a screen recording with an AI coding agent?
Yes — when a bug only shows up in motion, a still won't cut it. Maxisnap records a region or the full screen (Ctrl+Alt+3 / Ctrl+Alt+4), lets you trim the dead air, and exports an AI-optimized MP4 — roughly 1280 px, 24 fps, mono audio — engineered to be small enough to upload fast into tools that accept video. It's the same "right-sized for the model" idea, applied to video. Full details live on the screen recording for AI coding page.
Screenshots for AI coding — FAQ
Does Maxisnap work with Claude Code, Codex CLI, and Gemini CLI?
Yes. Any screenshot can be exported as a one-click, right-sized AI-optimized copy, which you then add by whichever route each tool documents; a recording exports as an AI-optimized MP4 for tools that accept video. Maxisnap produces the file; it doesn't automate or paste into the agent for you.
What size are the AI-optimized images, and can I change it?
The default maximum width is 1600 px, and it's configurable in Settings. That width is chosen so the image stays legible to the model without spending tokens on 4K resolution it downsamples anyway.
Why not just paste a raw screenshot into the terminal?
A full-resolution capture is usually more image than the job needs. Anthropic's vision guide, read in October 2026, says images above a model's size limits are downscaled before processing, and suggests downsampling images before sending them to control token costs when you do not need the extra fidelity.
What if pasting a screenshot does nothing in my terminal?
If pasting an image does nothing in your terminal, use a saved file instead. Maxisnap gives you a saved, correctly-sized image you can drag into the terminal or reference by file path.
Does it run on Windows?
Yes — Maxisnap is Windows-native (Windows 10 1903+ and Windows 11). A Mac version is also available, for macOS 13 or later.
Give your AI coding assistant a picture worth sending.
Capture, remove sensitive content, and inspect an AI-optimized copy before sharing — free on Windows.
Related: screen recording for AI coding · paste a screenshot into Claude Code · screen recorder + screenshot tool · blur a screenshot · all features
Maxisnap is not affiliated with the other products named on this page. They are described as their official sites presented them in October 2026; check each official site for current features, pricing and platforms.