Curio AI /// Enterprise & SaaS Platform | Product Design

Curio AI Platform

Curio AI is platform that connects to cloud and on-premise storage and uses computer vision, speech-to-text, OCR, NLP, and other ML models to automatically generate rich, second-by-second metadata across video, audio, images, and documents. Users can search for people, objects, logos/brands, locations, emotions, technical cues, transcribed speech (50+ languages), and more from a single interface or API—turning massive, previously unsearchable archives into instantly discoverable, monetizable assets. Originally developed at GrayMeta and later acquired by Wasabi Technologies (announced January 2024) to power AI-enhanced intelligent storage for media & entertainment.

Flowchart illustrating the process from stored media to searchable intelligence, including sections on user storage, ingest, AI/ML processing pipeline, metadata index, web UI and API, and downstream consumers.
A black background with vibrant green text saying: 'INGEST -> ENRICH -> INDEX -> DISCOVER -> USE'

The challenge

Media organizations sit on terabytes (or exabytes) of footage that is effectively dark without detailed metadata. Video editors waste hours scrubbing timelines or hunting through folders for the exact clip of a player celebrating, a logo appearance, or a specific spoken phrase. Manual metadata taggers face endless, error-prone work that cannot scale. Asset managers struggle to surface and monetize archival content for licensing, personalization, or secondary revenue.

Focusing on the user

Existing tools were either too shallow, required content to leave the customer’s environment, or forced users to bounce between systems. We conducted a deep dive into core personas:

Video Editors / Producers – Need frame-accurate search and direct pull into timelines without breaking creative flow.

Metadata Taggers / Archivists – Want automation that still allows human review and enrichment.

Asset Managers / Monetization Leads – Care about discoverability, rights, brand/logo tracking, and turning archives into revenue.

Technical / Storage Admins – Require secure connections to existing cloud/on-prem storage, API access, and controllable processing.Research included interviews, workflow mapping, and observation of real editing and archive sessions.

Key insights: search must feel instantaneous and natural-language-friendly; results need rich previews and filters; the system had to work across heterogeneous storage without forcing migration; and trust in AI accuracy (with easy override) was non-negotiable. This informed a unified search experience, progressive enrichment pipelines, and persona-specific views that keep every user focused on high-value work instead of hunting.

Close-up view of a computer screen displaying a web page titled 'Curio', showing search results for a video titled 'A man and a woman in a car wearing sunglasses'. The screen includes various icons, categories, and analysis tools, with a partial view of a woman and a man smiling and wearing sunglasses in a parked car.
Crowd of people at a convention or expo, gathered around a booth with a large blue and green sign that says "FIND IT ALL" and "curio.app." Several individuals are engaging with displays and technology, with some wearing lanyards and backpacks.
A laptop on a wooden table displaying a website or application with a photo of two women on the top left, and colorful horizontal charts or timelines below.

SXSW Booth

Curio AI SaaS

A digital style guide titled 'Curio style guide' displaying sections for typography, colors, icons, buttons, menus, forms, modals, and thumbnails, with color palettes, icon lists, and UI element examples.