This mime type lookup answers two directions of one question. Type an extension and read the media type. Type a media type and read the extensions in this list. Each hit also has one sentence of meaning. The table is in the page script. It is not the IANA media-type registry, and it is not a file sniffer.
The match is exact after a little normalization. A leading dot is optional. Case does not matter, so JPG, .jpeg, and image/JPEG hit the same row. .jpg and .jpeg are both image/jpeg. .htm and .html are both text/html. A media type must include the slash. json is the extension. application/json is the type. A fragment such as image is not a search across every image row.
The curated set covers documents, scripts, styles, JSON, XML, PDF, form bodies, the image types this site already talks about, common web fonts, zip, gzip, WebAssembly, and application/octet-stream. Form types and octet-stream have no extension in the list, so the extension cell says none. .js and .mjs are listed as text/javascript, which is the type HTML uses. Older names that are not separate rows are unknown here rather than guessed.
A value that is not in the table is an error: it is not in this curated list, and no type was invented for it. Vendor trees and one-off suffixes are omitted on purpose. SVG is in the list as image/svg+xml because that type is registered. The image tools on this site still refuse SVG files. A lookup is not an invitation to upload one.
Empty input is an error. The limit is 200 characters, which is longer than any name in the table. The result cells are text.
webp returns image/webp. image/jpeg returns .jpg, .jpeg. HTML returns text/html with both document extensions. application/wasm returns .wasm. application/x-www-form-urlencoded returns a row with no extension. application/vnd.example and .heic are not in the list.
webpThat is the whole interaction. There is no second page of history for the type, and there is no button that downloads a registry.
The page will not sniff bytes, will not choose a type for a picture you drop on it, and will not map every archive suffix. Zip and gzip are here. A tar suffix is not, because this list does not add an unregistered type to cover it. Octet-stream is here as the generic bytes type, and no extension points at it, so a random suffix does not collapse into that row.
Nothing you type is stored. Clear removes the query and the row. The next lookup reads the same fixed table.
Looks up a curated local list, not the IANA registry. .jpg and .jpeg are image/jpeg. .htm and .html are text/html. A value that is not in the list is reported as unknown. Nothing is fetched and no file is uploaded. Limit: 200 characters.
No. The list is a curated set of common web types that ships with the page. Missing values are reported as not in the list.
No. Both map to image/jpeg. htm and html both map to text/html.
No. You type an extension or a media type. Nothing is uploaded and nothing is fetched.
The page does not invent types. If the string is not in the curated list, the status says so.