Reinforcement learning environments in which live data streams in on a clock while an AI agent works, covering personal agents, industry operations and science and compute, for training models to reason over changing signals.
Perit
Training data made to order for AI labs: speech in nine languages, video, image and text collection, annotation on its own Foundry platform, and reinforcement-learning environments and benchmarks.
- Artificial Intelligence - AI
- Machine Learning
- Speech Recognition
- Robotics
TrustList summary
Not yet independently verified. Dated from the Y Combinator launch post of 14 November 2026; the maker's site gives no launch date, headquarters or prices, and its delivery figures are the maker's own, with the site and the launch post giving different totals. We will update this when it can be confirmed, and remove this note.
When a model team needs a few thousand hours of Korean or Hindi speech recorded and aligned to a specification, or first-person video of hands handling objects, someone has to recruit the speakers, run the sessions and check the labels. Perit, operated by Zyno AI Inc and backed by Y Combinator, sells that work to AI labs.
Speech in nine languages, and more than speech
Its main line is speech data in German, French, Italian, Spanish, Portuguese, Hindi, Chinese, Korean and Japanese. It also collects video, images and text to a customer's specification and annotates them on its own platform, Foundry. A newer line it calls Physical Intelligence covers egocentric video, depth data, teleoperation and gripper demonstrations for robotics teams; the company says part of this has shipped and part is still being scoped.
Environments and benchmarks
Beyond raw data, Perit builds reinforcement-learning environments and evaluation sets, including a speech-to-speech benchmark across ten sectors and a leaderboard for word alignment. It stresses that no frontier lab sits on its cap table, a point aimed at buyers who worry about a supplier also serving a rival.
How a project runs
There are no list prices. The site promises a quote within 48 hours, a sample episode within five days and a pilot batch with a quality report within two weeks, then weekly deliveries. Licences are non-exclusive by default, with exclusive rights priced separately.
Figures the maker gives
The site cites more than 4,000 hours of audio delivered, 650 transcribers and aligners, 50 hours produced each working day and more than 95% accuracy on quality samples. The Y Combinator launch post gives larger totals, more than 6,000 hours of speech and more than 10,000 hours of egocentric data, so the two pages do not agree.
Key facts
Maker: Zyno AI Inc (Perit.AI) Founders: Kshitij Dixit (CEO) and Poorva Dixit (CTO) Launched: Y Combinator launch post, 14 November 2026 Languages: German, French, Italian, Spanish, Portuguese, Hindi, Chinese, Korean, Japanese Pricing: quote on request Website: perit.ai
Categories & features
- Artificial Intelligence - AI
- Machine Learning
- Speech Recognition
- Robotics
Request a demo or quote from Perit
Reviews
Questions & answers
No questions yet. Be the first to ask about Perit.
Perit alternatives
Similar listings buyers compare against this one.
Small on-device AI models for speech, text and vision, delivered as SDKs for mobile, web and Node
Optimisation platform that benchmarks prompts, models and agents on each team's own data, with tracing, for AI product teams

Microsoft service for building, training and deploying machine learning models on premises or in the cloud
Automated machine learning platform for building, deploying and governing AI models

AI size and fit recommendations, virtual try-on and device risk signals for fashion retailers
Computer vision toolkit for managing datasets, annotating images and training models

Ad technology that uses machine learning to place and analyse ads for web publishers and brands
More on TrustList
Everything here links back to the same verified catalogue. Pick your next stop.
- Write a reviewHelp other buyers decide
- CompaniesAgencies, consultancies and IT service providers, ranked by verified reviews.
- ArticlesGuides, comparisons and research from the editorial desk and the community.
- AwardsAnnual recognition decided by verified reviews and an independent jury.
- LaunchesNew products and releases, voted up by the community every day.
- AI ModelsBenchmark scores and community ratings for every major model.
- RequestsBuyers describe what they need; vendors respond directly.
- PeopleReviewers, authors and makers with public profiles.
- ComparePut up to four listings side by side before you shortlist.




