# Clipboard Inspector > Clipboard viewer: every MIME type — text/plain, text/html, image/png, vendor formats. Raw HTML, sizes, hex. Local only, never uploaded, gone when you leave. - URL: https://bytescope.dev/en/clipboard-inspector/ - Locale: en · Also: [ja](https://bytescope.dev/ja/clipboard-inspector.md) · [zh-Hant](https://bytescope.dev/zh-Hant/clipboard-inspector.md) - Runs: Nothing is uploaded or logged. The snapshot lives in this tab and is gone when you leave. - Requires: Any modern browser. How many clipboard formats a page may see differs per browser and the page reports exactly what it received. - Updated: 2026-09-08 ## What it does This page answers a question developers actually type: what is in my clipboard? Not a guess. After you click, it asks `navigator.clipboard.read()` for every MIME flavour the browser will hand over — `text/plain`, `text/html`, `text/uri-list`, `image/png`, `image/svg+xml`, and the vendor ones like `web application/x-vnd…`. Each row is a type, a size in bytes, and a viewer that fits. Nothing is uploaded, nothing is logged, and leaving the page drops it. ### All the MIME types, not a textarea A paste into a form only keeps what that form asked for. The clipboard itself often holds several representations of the same copy at once. This page lists every one. Hex uses the same language as the hex editor: offset, 16 bytes per row, ASCII on the side. Images show type, bytes, and pixel size when `createImageBitmap` can read them. A flavour whose `getType()` rejects is marked unavailable; the rest of the table still renders. ### The HTML cell is the point Paste from Word or Notion and the hidden passenger is `text/html` — inline styles, `mso-` junk, converted lists. This page shows that source as text, escaped so it cannot become DOM. There is no `dangerouslySetInnerHTML`, no `eval`, no injection. If you only wanted the words, `text/plain` is the other row; the HTML row is why the paste looked wrong. ### Denied is not empty, and unknown is not denied `clipboard.read()` must run from a click or a paste. This page never reads on load. `NotAllowedError` means the browser refused or still wants a gesture — not that the clipboard is empty. An empty result is a different state. `navigator.permissions.query({ name: "clipboard-read" })` often throws; a throw is shown as `unknown`, never as `denied`. Some browsers only implement `readText()`. That is reported as "this browser only returns plain text" — not "the clipboard only had plain text". ### Two paths, because they disagree The Async Clipboard API and a `paste` event's `clipboardData.items` are not the same pipe. Chrome may sanitise HTML on `read()` and leave a `web text/html` unsanitized twin; a paste event can still see types `read()` hid, or the other way around. The paste well on this page is a comparison, not a second toy. Both results stay in this tab and vanish when you leave. ## Limits Does not show clipboard history or watch the clipboard in the background; it inspects what you paste or read on demand. ## Machine output Copy JSON copies an envelope `{ tool, url, generatedAt, userAgentHint, result }`. `tool` is `"clipboard-inspector"`; `result` is shape only — MIME types, byte lengths, item counts, image pixel size. Clipboard payload is omitted. - `clipboardRead` (boolean | null) — `null` = not probed yet; `false` = probed, `navigator.clipboard.read` missing; `true` = present. - `clipboardReadText` (boolean | null) — `null` = not probed yet; `false` = probed, `readText` missing; `true` = present. - `permission` (string | null) — `"granted"` | `"denied"` | `"prompt"` | `"unknown"`. `null` = permission query not asked. - `asyncSnapshot` (object | null) — Async Clipboard `read()` / `readText()` snapshot. `null` = no successful read. When present: `source` (`"read"` | `"readText"`), `textOnly`, `itemCount`, `items` (each item has `types`: mimeType, kind, unsanitized, byteLength, imageWidth, imageHeight). `byteLength` is `null` when that flavour was listed but not read; image dimensions are `null` when kind is not image or size did not load. - `pasteSnapshot` (object | null) — `paste` event snapshot. `null` = no successful paste. Same inner shape as `asyncSnapshot`, with `source` `"paste"`. ## FAQ ### Do you upload or store what is on my clipboard? No. The bytes never leave this tab. There is no server round-trip, no analytics payload with the contents, no log. Closing or navigating away drops the snapshot. The privacy line on the page is the whole policy. ### Why do I have to click? Why not read automatically? The Async Clipboard API requires a user gesture, and this page would not auto-read even if it did not. Clipboard contents are secrets: passwords, tokens, unsent mail. A click is the consent. The paste well is the other consent — you pasted. ### I clicked and it said permission denied. Is my clipboard empty? No. `NotAllowedError` is a refusal (or a missing gesture), not an empty clipboard. Empty is a separate result: `read()` came back with no items, or the browser threw `NotFoundError`. This page will not fold those together. ### The paste well and the Read button show different types. Which is real? Both, for their own API. `navigator.clipboard.read()` and `event.clipboardData` are different paths, and browsers sanitise them differently — especially `text/html`. The mismatch is the information. Neither path is rewritten into the other. ### This browser only shows text/plain. Is that all that was copied? Not necessarily. Some browsers only expose `clipboard.readText()`, which returns Unicode text and no other MIME types. This page says so in those words. It will not claim the clipboard "only contained plain text" when the other flavours were never offered. ### Is it safe to look at HTML from the clipboard here? The HTML is escaped and shown as text. It is not parsed into elements, not run as script, and not assigned with `innerHTML`. Treat it as untrusted input anyway — that is why it is a viewer, not a preview of the rendered page. ## Related tools - [Hex Viewer / Editor](https://bytescope.dev/en/hex-editor.md) — Same offset / 16-byte / ASCII language, for a file instead of a clipboard flavour. - [File Type Identifier](https://bytescope.dev/en/file-type.md) — Not sure what a binary flavour is? Identify it from magic bytes. - [Image ↔ Base64 Converter](https://bytescope.dev/en/image-base64.md) — Turn a clipboard PNG into a data URI, or the other way around.