KMS Web Clipper (clip web pages and screenshots to the Vault)
Save a web page by URL, or the text visible in a screenshot, straight into your KMS vault as a clean Markdown note with its source, timestamp and tags recorded. Pages are fetched and extracted for you; screenshots go through an AI read that costs a small amount. KMS must be enabled.
On this page
What it is
Status: shipped. Clip any web page by URL — or a screenshot by OCR — into the Vault as a Markdown note. Zero copy-paste friction.
The Web Clipper lets you save web content directly into your Vault's Web Clips/ folder as a clean Markdown note with YAML frontmatter (source URL, timestamp, tags). It fetches the page, extracts the article with Mozilla's Readability library, converts it to Markdown, and creates a new note — all in one step.
For screenshots, it uses AI-powered OCR (Anthropic vision) to extract visible text from the captured image and saves that as a Vault note. This incurs a small API cost (one Anthropic API call).
Requires KMS enabled. All entry points are gated behind the nothariEnabled setting. If KMS is off, the Quick Launch tab is hidden, the CLI route returns 403, and the screenshot button doesn't appear.
Where to find it
The fastest path is Quick Launch — press your Quick Launch shortcut and switch to the Web Clip tab, whose URL field auto-detects a link already on your clipboard. The Web Clip tab is not pinned by default: add it once from the strip's + button or Settings → Quick Launch tabs, and it stays. A screenshot button is available from the same place, and there is a command-line route for agents.
How it behaves
How to use it
1. Quick Launch tab (fastest for URLs)
Open Quick Launch (Ctrl+Space or your configured shortcut) → switch to the Web Clip tab (scissors icon). If you have never pinned it, there is no Web Clip tab yet: open the strip's + button (or Settings → Quick Launch tabs) and pin Web Clip first — it is not one of the tabs that ship pinned. The URL field auto-detects if your clipboard contains an http:// or https:// URL and pre-fills it. Optionally add comma-separated tags — letters, numbers, -, _ and / only, up to 10 tags; a leading # and spaces are cleaned up for you. Press Enter or click Clip page.
The modal closes immediately (fire-and-forget). The clip happens in the background. If it fails (unreachable URL, non-HTML page, no vault configured), a toast appears in the main window.
2. CLI route (for automation and agents)
POST /kms/clip
Content-Type: application/json
Authorization: Bearer <token>
{
"url": "https://example.com/article",
"tags": ["research", "ai"],
"title": "Custom title (optional)",
"vaultId": "optional-vault-id"
}
Returns 201 with the saved note nested under a data section: { ok: true, data: { note: { id, vaultId, relativePath, ... } } }. A repeated identical request answers with the same shape plus idempotent: true, so the page is filed once, never twice.
Tags follow the same rule: letters, numbers, -, _ and / only, up to 10 tags.
Agents can call this to save research material to the Vault during a session.
3. Screenshot "Save to vault" (OCR)
Capture a screenshot with Omniscio's snip tool (Ctrl+Alt+S or your configured shortcut). In the annotation editor, click Save to vault (only visible when KMS is enabled). The button shows "Extracting..." while the AI reads the screenshot.
The OCR extracts all visible text and generates a short title. The result is saved as a Vault note in Web Clips/ with tags screenshot and ocr, and a screenshot://<recordId> source URL for provenance.
What the clipped note looks like
---
source: https://example.com/article
clipped: 2026-07-17T19:30:00.000Z
tags:
- research
- ai
---
# Article Title
*Author Name* | [Source](https://example.com/article)
The article content in clean Markdown...
The note file lands at Web Clips/<title> <YYYY-MM-DD HHmm>.md. If a file with that name already exists, a numeric suffix is added automatically.
Limits and safety
- URL validation: only
http://andhttps://URLs are accepted. Private/local URLs (localhost, 10.x, 192.168.x, etc.) are rejected for SSRF safety. - Response size: 10 MB max. Pages larger than that are rejected.
- Note body: 5 MB max (inherited from the note creation service).
- Fetch timeout: 15 seconds to start responding. The timer stops as soon as the page answers — once its response headers arrive, not when the download finishes — so the download itself is bounded by the 10 MB size cap rather than by the timer.
- Content type: only
text/htmlandapplication/xhtmlpages are clipped. PDFs, images, and other content types are rejected with a clear error message. - Readability fallback: if the article extractor can't parse the page (JavaScript-heavy SPAs, paywalled content), the clipper falls back to plain text extraction and marks the note as a rough copy.
Troubleshooting
| Symptom | Cause | Fix |
|---|---|---|
| "Install The Vault from Marketplace → Inbuilt to use this tab" | The Vault is disabled | Install The Vault from Marketplace → Inbuilt, then set a vault folder |
| "The URL does not point to an HTML page" | Non-HTML content type | The clipper only handles HTML pages |
| "Cannot clip from a private or local URL" | SSRF protection | Use a public URL |
| "The page is too large to clip" | Response > 10 MB | Try a lighter page or clip manually |
| "<site> would not let us save that page (it answered with error 401/403)" | The site blocked the clipper — a sign-in wall, paywall or bot check | Not fixable from here; copy the article text into a note instead |
| "<site> no longer has that page (it answered with error 404/410)" | The link is dead | Check the link, or clip the live version |
| "<site> had a problem serving that page (it answered with error 5xx)" | The site is having an outage | Try again later, or copy the text into a note |
| "<site> refused the request to save that page (it answered with error 4xx)" | The site rejected the request for another reason | Try again later, or copy the text into a note |
| "This page did not have readable text we could save" | JavaScript-rendered page or a sign-in wall, so nothing was extractable | Copy the text into a note instead |
| A clipped note opens with "Rough copy" | Article extraction failed, so the page's raw text was saved | Read it as a fallback copy; the extraction: plain-text header records why |
| "Could not clip that page" (toast) | Fetch failed or no vault | Check the URL is reachable and you have a vault configured |
| "Couldn't extract text from this screenshot" | OCR API error | Check your API key is valid and you have credits |
Related
The vault the clips land in is described on the KMS page, and importing a whole existing note collection rather than clipping page by page is the KMS Obsidian Import Wizard.
Last verified 2026-09-28