Velvet partners with AI labs and enterprises to build and evaluate datasets that improve interactive AI capabilities. Their platform lets users submit dataset samples—including text, image, and video—for review, and connects creators with payment options via Stripe or Wise to monetize their contributions. This service streamlines the creation of high‑quality training data and evaluation benchmarks for advanced AI systems.
Funding
Funding not disclosed
Founders
Product
Problem
AI research labs and enterprises often struggle to obtain high-quality, domain‑specific datasets and evaluation suites needed to train and benchmark interactive AI systems. Existing data marketplaces lack focused curation for multimodal interactive use cases, leading to costly data collection and limited evaluation coverage.
Solution
Velvet provides a collaborative platform that links AI labs and enterprises with data owners to create specialized multimodal datasets and evaluation suites. Data owners upload sample text, image, or video assets through a guided portal, where Velvet’s team assesses suitability for interactive AI applications. Approved contributions are matched with lab or corporate requests, and payment is handled automatically via integrated Stripe or Wise accounts. Contributors can track pending and paid earnings through a personal dashboard, while labs receive curated, ready‑to‑use datasets and evaluation tools that accelerate model development and benchmarking.
Target Audience
Primary customers are AI research labs and enterprise AI teams seeking curated multimodal datasets and evaluation suites, as well as data owners and organizations looking to monetize their proprietary content.
Features
- Multimodal submission interface supporting text, image, and video samples with volume and metadata fields
- Expert review workflow that evaluates dataset relevance and quality for interactive AI tasks
- Built‑in payment integration with Stripe and Wise for seamless compensation and earnings tracking
- Contributor dashboard displaying pending, paid, and total earnings in real time
- Secure cloud storage links (e.g., Google Drive, Dropbox) for sample assets to ensure data integrity
- Matching system that connects approved datasets with AI lab and enterprise project requirements