Video & subtitles · YouTube Thumbnail Downloader & Metadata Viewer
How oEmbed returns a video's public title, channel and embed code
· How it works
youtube oembed metadata
There is a standard, key-free way to ask a video site for a link's public title, author and embed HTML. This post explains oEmbed, what a request and response look like and how a tool can show metadata without scraping the page.
Title and channel without loading the page — the question that oEmbed was designed to answer
A blogger can learn a public video’s current title and channel without loading the watch page. ToolAcre asks YouTube’s oEmbed endpoint for a structured JSON record only after the visitor explicitly presses Fetch. That sequence also prevents a failed metadata lookup from being confused with a malformed URL, since the two stages have different evidence.
The link is parsed first in the browser. That preflight validates and extracts the ID without network access, keeping malformed input away from both the thumbnail host and metadata endpoint. The endpoint supplies a narrow public representation, not a general-purpose data feed or an authorization mechanism.
What oEmbed is — a small standard where a site publishes an endpoint that answers 'describe this URL' with JSON
oEmbed is a request-and-response convention for describing a public URL. Here, YouTube provides an endpoint that accepts the video URL and returns JSON fields intended for embedding and attribution. The exact query is useful in a Network trace because it shows which canonical resource was requested without exposing account credentials.
ToolAcre treats that endpoint as a ceiling rather than a doorway to scraping. It does not inspect watch-page HTML, internal youtubei services, signatures or anything that a signed-out public request cannot reach. A field can be present in raw JSON and still be intentionally omitted from the application contract, so normalization is a product decision. That keeps downstream notes from depending on accidental upstream additions.
The request — the video URL and a format parameter, nothing more, sent to youtube.com
The request goes to https://www.youtube.com/oembed with url set to a canonical https://www.youtube.com/watch?v=ID value and format set to json. No API key, login cookie or account token is added. An empty result is consequently a boundary observation, not evidence that the identifier was invalid or that protected data should be retried.
The browser sends an anonymous GET with credentials omitted, no referrer, no-store caching and followed redirects. Google receives it and its Origin header directly; ToolAcre operates no server or proxy in that path. This keeps exports stable when upstream payloads contain extra fields that the tool does not promise to preserve.
The normalized response exposes metadata fields, not oEmbed HTML
The outline promised displayed embed HTML, but normalizeOembed deliberately leaves that field out. The normalized object contains title, channel, channel URL, thumbnail URL and dimensions, player width and height, provider and the video ID. Classifying failures separately helps a user decide whether to retry, inspect the response, or choose another permitted source.
Every value is defended: missing text gets a plain fallback, dimensions must be positive, and links must be HTTPS. This makes exported JSON predictable without implying that every raw oEmbed property is rendered. The request can be inspected in DevTools, but the browser record does not imply storage by ToolAcre or local-only processing.
Why no key or login is needed — the data is already public on the page; oEmbed just makes it structured
No key is required because this is signed-out public metadata, but public does not mean universal. Private, deleted and age-restricted videos do not provide the record, while an owner who disables embedding can cause an authorization response. A publisher reviewing a failed request should therefore check embed availability and video visibility rather than assuming the eleven-character ID was parsed incorrectly.
The absence boundary is explicit: no views, likes, description, tags, duration, publish date or captions are synthesized. A key-free endpoint does not grant access beyond what YouTube elects to publish. If an editorial spreadsheet needs those omitted fields, its author must use another permitted source and label that source instead of presenting blanks as oEmbed data.
Worked example: mapping oEmbed JSON into ToolAcre’s normalized record
For one successful payload, title maps to title, author_name to channel, author_url to channelUrl, thumbnail fields to a nested image description, and width and height to player dimensions. provider_name becomes provider. This mapping gives exports stable application names even though the upstream convention uses author terminology for a YouTube channel.
Raw HTML may be present upstream, but the ToolAcre result and exports do not expose it. The app separately builds its own responsive privacy-enhanced iframe from the validated ID, avoiding accidental trust in endpoint markup. Consumers receive facts they can quote or serialize, while executable embed markup remains outside the normalized metadata contract.
What this does not cover — view counts, descriptions, comments, captions and anything behind a login
A metadata attempt can fail before a response because of offline mode, blockers or proxies; it can fail through an HTTP status; or it can return a success status with invalid JSON. Each route receives a different message. A transport error suggests checking the browser environment, whereas invalid JSON identifies a response that arrived but could not satisfy the endpoint’s promised format.
Thumbnail probing is independent and may see 404, another error, a 200 placeholder or undecodable data. Neither branch downloads video nor bypasses restrictions for private, deleted or age-restricted content. Keeping the branches independent lets an editor retain a verified poster even if title attribution is temporarily unavailable, without inventing the missing metadata.
Structured public metadata without displaying the endpoint’s embed HTML
The result is structured public metadata, not a copy of arbitrary oEmbed HTML. ToolAcre can export the normalized facts as JSON, Markdown, CSV or text and can build known links from the validated identifier. The choice of export changes presentation, not scope: every format reflects the same title, channel, image, player, provider and ID record.
After Fetch, traffic consists of direct i.ytimg.com image requests and the one canonical oEmbed request. Requests omit credentials and referrer, use no-store and followed redirects, and remain visible for verification in the browser Network panel. That trace proves the browser contacted Google directly; it does not show ToolAcre storing a lookup, nor does it make the operation local.