AI metadata that speaks your taxonomy
Generic AI labels create a second language your systems do not speak and your creative teams do not use. Imaginario adapts every output to your vocabulary, your fields and your schema.
Taxonomy and schema mapping delivers metadata in your language: taxonomy configured at onboarding, schema mapped to your MAM or DAM fields, shipped as JSON and CSV or straight into your systems.
Already powering video search and indexing for companies like
Every archive has a language. Imaginario learns yours.
Enterprise libraries run on controlled vocabularies, rights fields and legacy schemas. Generic AI tagging ignores all of it and outputs its own labels, which turns every indexing pass into a cleanup project for your metadata team.
Imaginario runs two engines side by side: a multimodal semantic engine that searches the content itself, and a structured metadata engine that writes human-readable, time-coded records your editors can inspect, correct and export. Both speak your taxonomy.
Your taxonomy, configured at onboarding
During onboarding we map your controlled vocabulary, naming conventions and required fields. From then on, every indexing pass outputs your terms, not generic labels, across your whole library.
Your uploaded content is always private and is never used to train generative models.
Mapped to the fields your systems expect
Metadata is adapted to your MAM or DAM metadata schema and passed back through native integrations or our integration partner: fields, types and vocabularies the way your system of record expects them.
Available today through our onboarding service, or as a standalone service for API and platform enterprise clients. No remapping projects, no cleanup passes, no second language.
Coming soon: taxonomy and schema mapping through the API.
Time-coded, readable, correctable
Transcripts, speakers, faces, chapters and shot-level descriptions, all time-coded to the second. Editors and librarians can inspect and correct records, so governance stays with your team.
Delivered as JSON responses and CSV files with custom fields.
Enrich the asset. And every second inside it.
Your metadata taxonomy is the floor, not the ceiling. Imaginario enriches metadata at two levels, the asset record your catalog runs on and the time-coded layer inside the video, so thin records become complete ones.
Enrich the record
- Synopsis, descriptions and summaries, short and long, multi-format
- Cast and key people, built from who actually appears
- Genre, mood and era
- Keywords in your vocabulary, not generic labels
- Content categories for ad placement and brand safety
Enrich every second
- Shot and scene descriptions with temporal context
- Transcripts, speakers and chapters
- Faces and named people, moment by moment
- Logos and on-screen text
- Actions, locations and sound
- Audiovisual language, moods and genres, your own taxonomies and metadata, and much more
Both levels ship in your taxonomy and your schema, as JSON responses and CSV files or straight into your MAM or DAM.
Generic labels create cleanup. Your taxonomy ships ready.
Three ways teams generate video metadata. Only one arrives in your language, in your fields, ready for your systems.
Imaginario AI
Generic AI tagging
Manual logging teams
Vocabulary
Your controlled vocabulary, mapped during onboarding
Generic pre-trained labels
Your terms, applied by hand
Upkeep
Outputs follow your taxonomy, no upkeep
Their vocabulary, remapping is your problem
Drifts between loggers and passes
What the AI actually understands
Every modality at shot and scene level, with temporal context behind every term
One model, one signal, generic labels
Whatever one person notices in one viewing
Schema fit
Mapped to the fields your MAM or DAM expects, passed back ready to use
Fixed vendor schema, one shape for every customer
Whatever the spreadsheet template says
Output formats
JSON responses and CSV files with custom fields
Vendor JSON only
Spreadsheets and sidecar documents
Time-coding
To the second, at shot and scene level
Frame or segment labels, no narrative context
Timecode ranges typed by hand
Backlog and archives
Entire archive in one pass, at 10 to 20% of runtime
Re-run and re-bill per model for every new field
About 3 hours of logging per content hour, per pass
When the taxonomy changes
Re-map once, outputs follow, no re-processing project
Retraining or re-processing, billed again
Another manual pass across the library
Integration / workflows
Native MAM, DAM, storage and NLE integrations, plus an integration partner
API-based, requires dev resources and monitoring
Copy and paste between systems
Associated cost
Included: platform from $89 per user per month, API from 1 to 3 cents per minute
Per minute per model, plus the cleanup time after
About $2 per minute of content, every pass
Plugs into your MAM and DAM in seconds
Metadata, clips and collections flow back into the systems your team already runs.
It compresses the ideation and creation process from days to minutes”
Sally A. Curtis, Agency Owner and Podcaster
“Imaginario AI makes it easier to go through a 2 hour movies. It does it in just half an hour. It’s the perfect time saver!”
Camila Mejia, Channel Manager, Cineverse Networks
“Imaginario AI has been a game changer for us. We can provide more content and save time doing so!”
Jennifer Neal, Formula Done
From generic labels to metadata your systems run on. All in one single platform and API.
Leading studios, broadcasters, production companies and corporate marketing teams use our system to identify shareable moments, clip and repurpose for social channels in seconds.
Make the most of your content library and engage your fans with ease.
Warner Bros. Discovery saw a
80%
time reduction in
multi-platform editing workflows
Cineverse located specific clips
75%
faster using our labelless AI search
Natural language search opens up more opportunities. Identify seasonal themes, pre-approved B-roll, age-restricted content for compliance in different markets, and much more.
Remove human biases, errors and typos. Search your footage confidently, knowing every second has been indexed to the same level of accuracy and granularity.
The fastest way to get
structure
from your entire video library
See what else you can do with Imaginario AI
Frequently asked questions
Can Imaginario use our existing taxonomy?
Yes. During onboarding we configure your controlled vocabulary, naming conventions and required fields, whether it is a DAM taxonomy, a MAM schema or a homegrown vocabulary, and from then on indexing outputs use your terms.
When the taxonomy changes, the mapping is updated once and outputs follow, with no re-processing project and no extra passes.
What is metadata schema mapping?
Schema mapping adapts AI-generated metadata to the structure your systems expect: the fields, types and vocabularies of your MAM, DAM or CMS.
Imaginario delivers time-coded, human-readable metadata mapped to those fields, so records land in your system of record ready to use, with nothing to sync or polish.
What formats can we export metadata in?
JSON responses and CSV files, with custom fields that follow your schema.
Through native integrations or our integration partner, metadata, clips and collections land directly in your MAM or DAM.
Is taxonomy and schema mapping available through the API?
Not yet, API access is coming soon. Today taxonomy and schema mapping is delivered through our onboarding service, or as a standalone service for API and platform enterprise clients.
Outputs ship as JSON responses and CSV files that follow your schema, or land directly in your MAM or DAM through integrations.
What tools support automated metadata enrichment?
MAM and DAM systems offer AI tagging add-ons, and hyperscalers offer one indexing API per model. Imaginario delivers automated enrichment as a service, at both levels: asset records like synopsis, summaries, cast, genres and content categories, and time-coded metadata down to shot and scene level.
Everything arrives in your taxonomy and your schema, so records land ready to use with nothing to remap.
What is video metadata?
Video metadata is the information that describes a video file and what happens inside it. Asset-level metadata covers the whole title: synopsis, cast, genre, keywords and rights fields. Time-coded metadata describes moments: shots, scenes, speakers, faces, logos and actions, each with a timestamp.
Imaginario generates both levels automatically and maps them to your vocabulary and fields.
How does AI video metadata improve searchability?
Every second of footage becomes findable by what happens in it: who appears, what is said, what is on screen and how it is framed, in plain language across the whole library.
Search stops depending on filenames and manually attached tags, which is how around 80% of a typical archive goes dark in the first place.
What is a metadata schema?
A metadata schema is the structure that defines which fields describe an asset and what each field can contain: title, synopsis, cast, rights, categories and so on. Every MAM, DAM and CMS has one, and mismatched schemas are why metadata gets retyped between systems.
Imaginario maps its output to the schema you already run, at asset level and time-coded level, so nothing needs retyping.
Get started today
Enjoy thousands of minutes of AI analysis, transcriptions, search, and auto-captions, plus much more.