Schedule a meeting with our OTT expert to grow your streaming business today!Book a MeetingContact Us
en$

What Is Video Metadata? A Guide for Streaming Catalogs

Video metadata is the information that describes each video. Here is which fields a streaming catalog needs on every title, what goes wrong without them, and how AI can fill them in for a person to approve.

By Usama Nisar

Cover: What Is Video Metadata? A Guide for Streaming Catalogs
On this page
View as Markdown

What is video metadata?

Video metadata is the information that describes a video: its title, description, tags, category, age rating, language, cast, subtitles and artwork, plus technical details such as its format and length. It is data about data. Viewers, search and apps read it to find a video, decide whether to watch it and play it correctly.

That is the short answer.

The idea is older than video. Wikipedia's article on metadata describes it as data that defines and describes other data, and points to the library card catalog: each card held a book's author, title and subject. A streaming catalog works the same way. The video is the book. The metadata is the card that lets someone find it.

This guide is about metadata for video content in a catalog, not about one file on your phone. If you run a library of tens or thousands of titles, the details below decide whether people find and watch them.

File metadata vs catalog metadata: what is the difference?

File metadata is technical and written by a machine; catalog metadata is descriptive and written by people or AI. The camera, phone or encoder records file metadata on its own: codec, resolution, frame rate, duration, creation date, sometimes location. Catalog metadata is what your team adds so a viewer can understand the title: what it is called, what it is about, who is in it, who it is suitable for.

Most pages about video metadata cover the first kind: how to view or edit a file's technical details. That matters for archiving and editing. It does not sell a title.

A viewer scrolling a TV home screen never sees a codec. They see artwork, a title, a short description and an age rating. Those are catalog fields, and nobody fills them in automatically unless you set something up to do it.

File metadata

Catalog metadata

Written by

Camera, phone or encoder

Your team, or AI your team reviews

Examples

Codec, resolution, duration, date

Title, description, tags, rating, language

Main use

Playback, editing, archiving

Search, browsing, choosing, ads

Who sees it

Mostly software

Viewers, on every screen

What metadata should every title in a streaming catalog have?

Every title in a streaming catalog should have at least ten video metadata fields: title, description, tags, category, maturity rating, spoken language, chapters, cast, subtitles and artwork. Each one has a job, and each one causes a specific problem when it is missing.

Field

Who uses it

What breaks when it is missing

Title

Viewer, search, TV app

The title cannot be found by name, or shows a file name such as "final_v3"

Description

Viewer, search

Viewers cannot tell what it is about and skip it

Tags

Search, recommendations, ads

Searches for a topic, a person or a theme miss it

Category

TV app rows, browsing

The title never appears in the right row or genre page

Maturity rating

Viewer, parental controls, ads

No age guidance; family profiles and ad targeting cannot rely on it

Spoken language

Viewer, player, search

The wrong audio track plays, or language filters miss it

Chapters

Viewer, player

Long videos have no way to jump to a section

Cast

Viewer, search

Searching a presenter's or actor's name finds nothing

Subtitles

Viewer, accessibility

Viewers who need or prefer captions leave

Artwork

TV app, viewer

Rows on a TV home screen show blank tiles or stretched images

Artwork deserves a note. A TV app usually needs several shapes for one title: a landscape tile, a portrait poster, a logo and a wide backdrop. One image cropped badly into all four is a common reason a home screen looks unfinished.

What goes wrong when metadata is missing or inconsistent?

Missing or inconsistent metadata makes titles hard to find, hard to choose and sometimes wrong to play. Nothing crashes. The catalog just performs worse, and it is hard to see why.

The usual failures look like this:

  • Search finds nothing. A viewer types a speaker's name or a topic. The title exists, but no field holds those words.

  • TV rows look empty. A row on the home screen shows grey tiles because a title has a landscape image but no portrait one.

  • No age rating. Parents cannot judge a title, and an ad partner cannot tell what is safe to place next to it.

  • The wrong audio language. A title recorded in Spanish is marked as English, so it plays in a language the viewer did not expect, or never shows up under Spanish.

  • Two names for one thing. One editor tags "yoga", another "Yoga Classes", a third "stretching". Filters and rows split the same content three ways.

Inconsistency is the quieter problem. It grows with the library. A catalog of 20 titles can be checked by eye. A catalog of 2,000 cannot.

Can AI write video metadata, and who checks it?

Yes, AI can write video metadata from the video itself, and a person should check it before it goes live. The useful setup is AI that drafts and a person who decides: the AI does the slow part, your team keeps control of what viewers see.

In Vodlix, this is Vodlix Intel. It writes each title's metadata from the video itself: title, description, tags, category, maturity rating and spoken language. Your team accepts each suggestion before it goes live.

Intel also watches and listens to each video to suggest chapters, cast, subtitles, dubbed audio in 14 languages and artwork up to 4K. Its Thumbnail Studio draws landscape, portrait, logo and hero backdrop artwork at 1K, 2K or 4K from a title's name, synopsis, tags and existing images, and holds it for your review. It uses the plan's monthly AI tokens.

How an AI metadata suggestion reaches viewers

  1. Start: Upload the video

    The title is added to your catalog.

  2. Step 2: Vodlix Intel suggests fields

    Title, description, tags, category, maturity rating, spoken language.

  3. Step 3: A person reviews it

    Someone on your team reads each suggestion.

  4. ?

    Decision: Is the suggestion right?

    Accept it as written, or change it first.

    No: edit the field, then accept

  5. Step 5: Accept the suggestion

    Only accepted fields are used.

  6. End: Live on every screen

    Viewers see the accepted metadata.

Not every part comes with every plan. AI subtitles come with Grow and up. AI dubbing and cast recognition come with Scale and up. Cast recognition finds the faces on screen, links them to artists already in your catalog and suggests names for new ones.

If your titles are already listed in The Movie Database (TMDB), there is a second route. Vodlix integrates with TMDB to fill in title data and artwork automatically, so a known title can start from an existing record instead of a blank form.

The guide video below shows Intel's suggestions on a single title, and how each one is accepted.

Vodlix Intel on a single title

Reviewing and accepting AI suggestions on the editor's Vodlix Intel tab.

How do you keep metadata right across a whole catalog and every screen?

You keep metadata right by holding one record per title, using a fixed list of categories and tags, and letting every app read from that same record. When the web, phone and TV apps each keep their own copy, they drift apart.

A few habits help at any size:

  1. Decide the list first. Fix your categories and your main tags before you upload in bulk. Write them down.

  2. Make the ten fields required. A title is not ready until each field in the table above is filled in or deliberately left empty.

  3. Review in batches. Accepting AI suggestions for a whole season or series at once is faster than editing title by title.

  4. Check on a TV. Artwork and descriptions that look fine in an admin screen can be cut off on a television.

In Vodlix, one catalog plays on the web, iOS, Android, Android TV, Fire TV, Apple TV, Roku, Samsung and LG TVs, with Chromecast and AirPlay casting. Which apps come with each plan is on the pricing page. The branded player on every screen shows a detail block with the title, poster, description and age rating, plus skip intro and chapters, so the fields you fill in once appear the same way everywhere. The metadata management feature page and the video content management platform overview show how the catalog is organized.

Start with your own catalog

The quickest way to see your gaps is to run your own titles through the process. Upload a few, let Vodlix Intel suggest the fields and review what it writes. You can start a 14-day free trial; nothing is charged until the trial ends.

FAQs

What are examples of video metadata?

Common examples are a video's title, description, tags, category, maturity rating, spoken language, chapters, cast, subtitles and artwork. Technical examples, written by the camera or encoder, include the codec, resolution, frame rate, duration and creation date.

What is structural metadata?

Structural metadata describes how the parts of something fit together, rather than what it is about. In a streaming catalog, that means things like which episodes belong to which season of a series, and where the chapters of a long video start and end.

How do I change the metadata on a video file?

For a single file, use the tool that holds it: the info or properties panel on your computer or phone, or your video editing software, can change details such as the title or date. For titles in a streaming catalog, change the metadata in the platform's content manager instead, so every app shows the update.

How do I check a video's metadata on an iPhone?

Open the video in the Photos app and open its info view, which shows basic file details such as the date and, when it was recorded, the location. For more technical detail, such as the codec, you need a video editing or file inspection app.

Does AI-written metadata go live without anyone checking it?

It should not. In Vodlix, Vodlix Intel suggests the title, description, tags, category, maturity rating and spoken language from the video itself, and each suggestion waits until someone on your team accepts it.

Can metadata be imported instead of written?

Yes, for titles that already have a public record. Vodlix integrates with The Movie Database (TMDB) to fill in title data and artwork automatically, which your team can then review and adjust.

Liked what you just read?

Subscribe to get the latest news, strategies, and insights on membership businesses delivered straight to your inbox.

By subscribing, you agree to receive occasional marketing emails from us. You can unsubscribe anytime with a single click.

Share this article

Latest Blog Posts

View All
Cover: Podcast App Development: Launch Your Own Branded Podcast App
Articles

Podcast App Development: Launch Your Own Branded Podcast App

How to get your own branded podcast and video podcast app on phones and TVs without building one, what each plan includes and costs as of November 2026, and how to earn from subscriptions, ads and free episodes.

Usama Nisar · Nov 17, 2026 · 13 min read