AI-Powered Metadata Generation AI-Powered Metadata Generation

AI-Powered Metadata Generation

Overview

AI-Powered Metadata Generation is part of a new AI Tools section in Zype CMS, which serves as the central hub for all AI-powered video automation. From AI Tools, you can enable and configure AI Metadata, AI Transcriptions, and AI Translation in one place.

The AI Tools hub — navigate here via AI Tools in the Zype CMS left nav. Enable AI Metadata, AI Transcriptions, and AI Translation from a single centralized page.  More AI Tools are on the way!

What’s New at a Glance

  • AI Generate Metadata button on every video’s Details page
  • New AI Tools section in the Zype CMS left navigation
  • Per-field wand icon (✨) for single-field AI generation next to every compatible field
  • Undo (↺) and Redo (↻) buttons for instant reversal of the last AI generation run
  • Bulk AI generation — select multiple videos in the Library and generate metadata for all at once
  • Review Updates modal with Update Warning, Missing Transcript detection, and AI acknowledgment
  • Brand Description & Editorial Guidelines — configure your voice once, apply to all generation
  • Reference Videos — up to 5 example videos to guide AI style and tone
  • Field Version History — up to 10 previous values stored per field, with per-field or full-session restore
  • Multi-language metadata output — generate metadata in any supported language
  • Auto-generation on upload, import, and MRSS feed sync
  • Usage & Billing dashboard for real-time token consumption tracking

AI Tools Hub

The AI Tools hub is the starting point for all AI features in Zype CMS. Navigate to it via the AI Tools link in the left navigation menu.

The hub contains cards for three AI capabilities:

  • AI Metadata: Automatically generate titles, descriptions, keywords, and genres from your video’s transcript.
  • AI Transcriptions: Generate highly accurate transcripts and subtitle files for your videos. Required for AI Metadata.
  • AI Translation: Translate transcripts into additional languages to expand your global audience reach.
ℹ️  Note: AI Metadata requires AI Transcriptions to be enabled. When you enable AI Transcriptions, AI Metadata can then be enabled. The hub status indicators will reflect both.

Setting Up AI Metadata

After enabling AI Metadata from the hub, configure it at AI Tools → Manage AI Metadata.

AI Transcription Settings

Because AI Metadata generation relies on transcript data, configure AI Transcriptions first:

  • Automatically AI Transcribe Videos: toggle ON to transcribe every new upload automatically.
  • Text Replacements: define custom substitutions to correct brand names, terminology, or syntax within transcripts (e.g., “Zy pe” → “Zype”).
AI Transcription Settings — configure the transcription model, auto-transcription toggle, and custom text replacements.

AI Metadata Settings

In AI Metadata Settings, you will find three configuration areas:

About Your Brand

Enter your Brand Description & Editorial Guidelines. This free-text field describes your brand, content categories, target audience, and editorial tone. The AI applies this context to every generation prompt. The more specific you are, the more on-brand the output will be.

AI Metadata Settings — the About Your Brand section and AI Metadata/Transcriptions status indicators. The blue info banner confirms that AI Transcriptions must be enabled before AI Metadata will activate.

Metadata Generation

Configure how the AI generates metadata:

  • Reference Videos: Add up to 5 videos from your library. The AI uses their existing metadata as style examples.
  • Field Toggles: Enable or disable generation per field: Titles, Friendly Titles, Descriptions, Short Descriptions, Keywords, Genres.
The Metadata Generation section — model selector, Reference Videos, per-field ON/OFF toggles, and the Usage & Billing section with a link to the Usage Dashboard.

Adding Reference Videos

Click + Add Reference Video to search your library and select an example video. The AI will use that video’s existing title, description, and keywords as a style reference for all generation.

The Add Reference Video modal — search your library and select up to 5 videos to use as style references for AI metadata generation.

Generating Metadata on a Single Video

Every video in your library has AI generation controls directly on its Details page. Open any video, navigate to the Details tab, and you will see:

The Video Details section. The AI Generate Metadata button runs generation across all enabled fields. Undo (↺) and Redo (↻) are to its left. The yellow wand icon next to each field generates AI content for that field only.

AI Generate Metadata Button

Clicking AI Generate Metadata generates suggestions for all enabled fields at once. Fields updated by the AI will reflect the new values immediately. Review each field, edit if needed, and click Save Changes to commit. Use the Undo button (↺) to instantly reverse the last generation run if needed.

Per-Field Wand Icon

Each AI-compatible field (Title, Description, Short Description, Genre, Additional Keywords) has a yellow wand icon next to it. Clicking the wand generates a suggestion for that field only — without touching any other field on the video. This is ideal for spot-updating individual fields.

Undo and Redo

  • Undo (↺): Instantly reverts all fields updated in the last AI generation run.
  • Redo (↻): Re-applies the last AI suggestion if you’ve pressed Undo (“Redo last AI suggestion”).
✅  Tip: For older rollbacks or per-field restoration, use Version History (the clock icon). Undo/Redo only cover the most recent run.

Bulk AI Metadata Generation

Generate metadata for multiple videos simultaneously from the Video Library.

Starting a Bulk Run

  1. Navigate to Library in Zype CMS.
  2. Select the checkboxes next to the videos you want to update.
  3. Click AI Generate in the bulk action toolbar.
The Video Library bulk action toolbar with the AI Generate button — visible when one or more videos are selected.

The Review Updates Modal

Before generation runs, the Review Updates modal summarizes the scope:

  • Videos Selected / Fields Selected / Total Updates: counts of scope.
  • New Generation (left panel): fields that are currently empty and will receive new values.
  • Update Warning (right panel): fields that already have values and will be overwritten. Review this carefully.
  • Missing Transcript Warning: shown when any selected video lacks a transcript. A “Proceed with Transcript Generation” button lets you generate transcripts first.
  • AI Acknowledgment: you must check the box confirming that AI-generated content may contain inaccuracies before proceeding.
The Review Updates modal. Left: New Generation fields. Right: Update Warning for fields with existing values. Below: Missing Transcript warning with Proceed option, Usage & Billing summary, and the AI acknowledgment checkbox.

 

After checking the AI acknowledgment, the Generate Metadata button activates. Click it to start the bulk generation job.

 

🚨  Important: Fields listed in the Update Warning panel will have their existing values overwritten. There is no bulk undo — use Version History on individual videos to restore specific fields after a bulk run.

 

Field Version History and Reverting AI Changes

Every AI-driven update to a metadata field is automatically tracked. You can view and restore any of the last 10 values per field at any time.

Opening Version History

On any video’s Details page, click the clock icon to the left of the AI Generate Metadata button (tooltip: “View Version History”). The Version History panel opens on the right side of the page, showing a timestamped list of past update sessions.

The Version History panel showing timestamped AI update sessions. Each entry shows the date, user, and number of field changes made in that session. Click the restore icon to revert.

Restoring a Previous Version

Click any session entry to open the Confirm Changes modal, which shows a Before / After comparison for every field changed in that session. From here:

  • Revert Change: restores a single field to its previous value.
  • Revert All Changes: restores all fields from that session at once.

Click Save Changes on the video to commit the revert.

The Confirm Changes modal — showing Before and After values for each field in the selected history entry. Use Revert Change for individual fields or Revert All Changes to restore everything.

 

ℹ️  Note: Version History stores up to 10 entries per field by default. This limit is configurable per site — contact your Account Manager to adjust it.

Auto-Generation Metadata in Upload, Import, and MRSS Workflows

AI Metadata generation can be triggered automatically as content enters Zype, so metadata is ready the moment a video is ingested.

Upload Workflow

On the Upload Video page, three AI toggles are available:

ToggleWhat it does
Enable AI TranscriptionTranscribes the video and generates SRT/VTT subtitles after upload completes.
Enable AI TranslationTranslates the generated transcript into configured target languages.
Auto Generate MetadataAfter transcription completes, automatically generates AI metadata using enabled field settings.

 

The Upload Video page with all three AI toggles — Enable AI Transcription, Enable AI Translation, and Auto Generate Metadata. Toggle all three ON to automate the full AI pipeline for every upload.

 

MRSS Feed Workflow

When configuring an MRSS source, the same AI toggles are available in the feed settings. Enable AI Metadata Generation to automatically generate metadata for every video synced from that feed.

MRSS feed settings showing Enable AI Metadata Generation, alongside Enable AI Transcription and Enable AI Translation toggles.

Pricing

As part of general availability, AI Metadata Generation is now available with usage-based pricing, measured per 1,000 tokens:

Token TypePrice Per 1000 TokensIncluded in Free-Tier
AI Metadata Input Tokens0.00126,000,000 tokens a month
AI Metadata Output Tokens0.0110800,000 tokens a month

All customers receive the monthly free tier shown above. Usage beyond the included free tier will be billed based on actual token consumption.

How token usage works

AI Metadata Generation is priced based on token usage. Tokens are small units of text processed by the AI model. Input tokens represent the transcript and context sent to the AI, while output tokens represent the metadata generated by the AI.

Related Articles

  • Getting Started: Enabling AI Metadata in Zype
  • AI Metadata Settings Reference
  • Using the AI Wand Button on a Video
  • Bulk AI Metadata Generation
  • Field Version History and Reverting AI Changes
  • AI Metadata in Upload, Import, and MRSS Workflows
  • AI-Powered Metadata Generation — FAQ