Contra Labs provides AI teams with professional creatives’ judgments, model evaluations, benchmark datasets, human-feedback training data, and recorded design workflows to improve and measure generative AI.
What they offerHuman data, Evals and benchmarks, Training data
Type of workCreative
Data sourcesBuilt from customer data, Synthetic / generated
Intelligence runs DesignArena, where people vote between AI-generated websites, games, images, video and audio, and sells model evaluations to AI labs. It says it shares users' prompts, outputs and votes with AI model developers, who may train on them, after taking steps to remove identifiers like usernames; its terms let it charge for that use.
What they offerEvals and benchmarks, Human data
Type of workCoding, Creative, Games, Enterprise workflows
Metaphi sells custom and reusable simulated work environments, training datasets, and agent evaluations to frontier AI labs and enterprise AI teams, helping them train and assess agents on realistic software, finance, and content-generation tasks.
What they offerRL environments, Training data, Evals and benchmarks
Type of workCoding, Enterprise workflows, Finance, Creative
Data sourcesLicensed real-world data, Expert or contractor-created
Moonlake provides world-model and simulation tools that turn real-world inputs or text into simulation-ready 3D assets and environments for robotics, autonomy, and AI research teams.
What they offerRL environments, Training data, Evals and benchmarks
Type of workRobotics and physical, Computer use, Creative
Protege licenses data from hundreds of partner organizations, curates it and de-identifies health records, and licenses the resulting datasets to AI developers for pre-training, post-training and evaluation across healthcare, video, audio and speech, and spatial and physical data. It also runs independent, double-blind evaluations of healthcare AI models on de-identified patient records.
What they offerTraining data, Evals and benchmarks
Type of workMedical, Creative, Robotics and physical, Multi-domain
Sama sells human-in-the-loop data annotation, validation, evaluation, and training-data services to enterprise AI teams building computer-vision, multimodal, and generative AI models.
What they offerTraining data, Evals and benchmarks, Data annotation, Human data
Type of workMulti-domain, Robotics and physical, Creative
Data sourcesBuilt from customer data, Expert or contractor-created
Sieve supplies AI labs and model developers with curated or custom-collected video, audio, image, and interaction data, plus annotations, evaluation sets, and interactive environments.
What they offerTraining data, Evals and benchmarks, RL environments, Human data
Type of workCreative, Computer use, Robotics and physical, Games, Multi-domain
Data sourcesLicensed real-world data, Expert or contractor-created
Taste Labs sells preference and reasoning data, rubrics, and evaluation environments to frontier AI labs for subjective design quality, and provides agent and app developers a Brand API and MCP tools to extract, search, and verify brand systems.
What they offerRL environments, Training data, Evals and benchmarks, Human data
TensorTest sells private evaluations, reinforcement-learning environments and expert-made training data to AI labs building creative agents, starting with video editing; its public TimelineBench benchmark scores agents on 56 editing jobs.
What they offerEvals and benchmarks, RL environments, Training data, Human data
Wirestock supplies AI labs with licensed, creator-made multimodal data, offering catalog datasets and custom collections for training AI models.
What they offerTraining data, Human data
Type of workCreative
Data sourcesExpert or contractor-created
Last checked
Initial directory snapshot · exported Nov 4, 2026. Original source check dates are retained. Automatic monitoring is not active yet. How we check information