Skip to main content
Generalhashgraph-online

separateweb-capture

Capture a URL into a full-page screenshot, cropped UI item PNGs, and a JSON manifest. Use when the user says `separateweb capture <url>`, asks to capture a website, or wants UI extraction assets without running the SeparateWeb web app.

Stars
336
Source
hashgraph-online/awesome-codex-plugins
Updated
2026-05-27
Slug
hashgraph-online--awesome-codex-plugins--separateweb-capture
View on GitHubRaw SKILL.md

// install — copy + paste into any project

mkdir -p .claude/skills && curl -fsSL https://raw.githubusercontent.com/hashgraph-online/awesome-codex-plugins/HEAD/plugins/AUN-PN/SeparateWeb/skills/separateweb-capture/SKILL.md -o .claude/skills/separateweb-capture.md

Drops the SKILL.md into .claude/skills/separateweb-capture.md. Works with Claude Code, Cursor, and any agent that loads SKILL.md files from .claude/skills/.

SeparateWeb Capture

Use this skill when the user asks:

separateweb capture <url>
npx separateweb-capture capture <url>
separateweb patch <path>

Use npx when the package is not linked globally. This is the safest default for Claude Code personal skills:

npx separateweb-capture capture <url>
npx separateweb-capture patch <path>

Use separateweb-capture with no capture command only to install the local skill payload:

npx separateweb-capture
npx separateweb-capture --target claude
npx separateweb-capture --target both

If separateweb is installed or linked globally, this shorthand is also valid:

separateweb capture <url>
separateweb patch <path>

If running from the plugin root, the local script is also valid:

node scripts/capture.mjs capture <url>
node scripts/capture.mjs patch <path>

Optional flags:

--out <dir>
--path <dir>
--width <px>
--height <px>
--max-pages <n>
--single
--all

Capture behavior:

  • capture https://example.com and capture https://example.com/ crawl same-origin paths by default.
  • capture https://example.com --single captures only the home page.
  • capture https://example.com/path captures only that page by default.
  • Use --single to force one page.
  • Use --all to force same-origin crawl from any start URL.
  • Captured items are grouped by type under items/<kind>/.

Examples:

npx separateweb-capture capture https://domain.com
npx separateweb-capture capture https://domain.com --single
npx separateweb-capture capture https://domain.com/about
npx separateweb-capture capture https://domain.com/about --all

Patch selection:

npx separateweb-capture patch /absolute/output/path
npx separateweb-capture select captures/<jobId>/manifest.json
npx separateweb-capture create captures/<jobId>/manifest.json --items 1,3,5 --path /absolute/output/path

Use patch <path> to set the default machine path where future capture outputs are written. Use create only when exporting selected items from an existing manifest.

Default output:

captures/<jobId>/site-manifest.json
captures/<jobId>/page-001-<slug>/full-page.png
captures/<jobId>/page-001-<slug>/manifest.json
captures/<jobId>/page-001-<slug>/items/<kind>/*.png

Single-page capture writes full-page.png, manifest.json, and items/<kind>/*.png directly under captures/<jobId>/.

After running, report:

  • Captured
  • Manifest
  • Blocks
  • exact error if capture failed

Do not start a Nuxt dev server. This plugin is script-only.