[Blog & guides](https://openscreenshot.app/blog/) · Developers & agents

# Take website screenshots from the command line

Use the OpenScreenShot CLI to save PNG screenshots with a fixed viewport, full-page capture, or binary output to stdout.

By [OpenScreenShot](https://github.com/pghqdev/OpenScreenShot)

The OpenScreenShot CLI captures a webpage as a PNG using a locally installed Chrome-compatible browser. It is a separate package from the browser extension. With Node.js, pnpm, and Chrome installed, run:

```
pnpm dlx openscreenshot shot https://example.com --out screenshot.png --full
```

The command starts a separate headless browser, navigates to the URL, writes the image, and closes the browser. It does not attach to the tabs or signed-in profile in your everyday browser.

## Set the viewport explicitly

For a viewport screenshot, omit `--full`. Width defaults to 1280 pixels and height defaults to 800 pixels. Set both when a layout needs a particular size:

```
pnpm dlx openscreenshot shot https://example.com --out desktop.png --width 1440 --height 900
pnpm dlx openscreenshot shot https://example.com --out narrow.png --width 390 --height 844
```

Width accepts integers from 200 to 3840; height accepts integers from 200 to 2160. A narrow viewport tests responsive layout at that width. It does not emulate a phone’s touch input, device pixel ratio, or mobile browser: the CLI uses a desktop user agent.

Add `--full` to capture beyond the viewport. The headless browser’s full-page capture differs from the extension’s scroll-and-stitch implementation; do not assume every dynamic page will look identical in both.

## Save a file or write to stdout

Without `--out`, the command writes `screenshot.png` in the current directory. Use an explicit filename to make artifacts easy to identify. Create any parent output directory before running the command.

`--out -` writes PNG bytes to stdout:

```
pnpm dlx openscreenshot shot https://example.com --out - > screenshot.png
```

Use binary-safe redirection. The CLI always produces PNG; naming the output `capture.jpg` or `capture.pdf` does not convert it. For annotated images or PDF output, use the [extension editor](https://openscreenshot.app/docs/#export).

## Redact before you save

Three flags cover data with black boxes in the page before the screenshot, so the unredacted pixels never reach the PNG. Repeat each flag for more targets:

```
pnpm dlx openscreenshot shot https://example.com --out safe.png \
  --redact "#account-menu" --redact-pattern email,secret --redact-rect 0,0,200,40
```

`--redact` covers each element that matches a CSS selector, with all of its children. `--redact-pattern` covers text and form values that match `email`, `phone` (9 to 15 digits), or `secret` (common API key and token shapes, and long hashes). `--redact-rect x,y,w,h` covers a rectangle in pixels of the output image; use it for canvas, images, and cross-origin iframes.

The capture fails with exit code 1 and writes no file if a selector matches nothing, a rectangle is outside the image, or the page moves a target during the capture. Patterns can miss text that is split across elements or drawn in an image, so look at the PNG before you share it.

## Resolve common failures

If Chrome is not found, install it or set `CHROME_PATH` to the browser executable. For example, on a Linux system where Chromium is installed at that path:

```
CHROME_PATH=/usr/bin/chromium pnpm dlx openscreenshot shot https://example.com --out screenshot.png
```

Navigation waits for `networkidle2` with a 30-second timeout. The current command has no custom wait, selector wait, cookie, or login options. A successful capture also does not prove that the application loaded correctly: inspect the PNG for error pages, loading states, and missing assets.

Exit code 0 indicates the capture command completed, 1 indicates a capture failure, and 2 indicates invalid usage or validated arguments. Use those codes when chaining commands, then review the image itself.

For repeatable artifacts, read [the CI capture guide](https://openscreenshot.app/blog/screenshots-for-ci/). For an agent-driven workflow, see [the MCP setup guide](https://openscreenshot.app/blog/screenshot-mcp-server/). The [CLI source](https://github.com/pghqdev/OpenScreenShot/tree/main/mcp/src) is the reference for the available flags.

## Keep going

[Read the documentation](https://openscreenshot.app/docs/), [browse all guides](https://openscreenshot.app/blog/), or [get help with a capture](https://openscreenshot.app/support/).

---

Canonical page: https://openscreenshot.app/blog/screenshot-cli/