Versos AI, the video training data infrastructure layer for library owners and AI labs, today announced a new capability built with NVIDIA NeMo that lets AI teams describe the video dataset they need in plain language to direct the Versos platform to find, evaluate and assemble matching content into a structured training dataset. For AI companies, this shortens a process that can otherwise require days of manual searching, review, and dataset preparation. The announcement was made from Saint John, New Brunswick, on September 9, 2026.
Dataset requests for training AI models often combine highly specific content needs with technical and licensing requirements including duration, resolution, format, language, production quality and permitted use. Fulfilling requests can require translating specification into multiple searches, manually inspecting large numbers of results, comparing assets, and organizing footage for delivery.
Versos new capability allows buyer to communicate requirements in single natural language conversation. System interprets prompt, translates into structured criteria and searches Versos scene and frame-level video intelligence. It then grades candidate footage against specification and assembles matching assets into dataset. For owners of premium video libraries, it helps content become more discoverable and usable to create stronger partnerships with AI companies.
"AI teams should be able to search for the data they need to train new models the same way they would describe it to a colleague," said Chris Keevill, co-founder and CEO of Versos AI. "The people building AI deserve to benefit from it just like everyone else, and our agentic interface makes that happen by streamlining the process for curating training data. NVIDIA technologies help us coordinate the work behind the experience, so the process is simpler for the user and more scalable for content owners."
Agents carry out all steps required to turn buyer request into usable collection of video data while keeping rights and provenance central. Workflow searches footage prepared for controlled AI licensing and maintains ownership, licensing and provenance information associated with content, shortening path to obtain structured datasets with documented permissions.
Versos built agentic curation workflow on NVIDIA and LangChain technologies at three levels: NVIDIA CUDA Toolkit accelerates inference inside Versos Video Library Intelligence Platform; NVIDIA Nemotron Ultra runs agents that interpret buyer request and carry out discovery and curation; LangChain provides agent scaffolding layer coordinating steps between plain-language request and platform search, metadata, grading and dataset-assembly functions. Architecture is model-agnostic by design, with same workflow able to run on other open-weight or frontier models where performance, cost or deployment requirements call for it.
About Versos AI Inc.
Versos AI is the AI licensing platform for premium video libraries. Versos AI enables studios, content owners, creators, and AI model builders to analyze, structure, license, and deliver video data for AI model training at scale. Through its Video Library Intelligence Platform, Versos AI indexes video at the frame level and transforms full libraries into searchable, AI-ready datasets for model training. With Versos AI, video libraries become monetizable datasets ready for use across the rapidly growing AI data ecosystem.
