Real-world data lab captures expert human demonstrations, builds RL environments from real workflows and evaluates agents with domain experts; first open benchmark on household manipulation due Q4 2026
SAN JOSE, Calif., Sept. 23, 2026 /PRNewswire/ — Realset AI (https://realset.ai), a real-world data lab that produces training data for LLMs, frontier models and embodied agents, today announced that Realset AI and Flatkey have raised $10 million in Series A funding. The funding will expand Realset’s capture network of real workplaces and studio environments, grow its pool of expert demonstrators and domain experts, and support open benchmarks that measure whether AI policies work outside the lab.
Why It Matters: The Internet Is Exhausted, the Physical World Is Not
Frontier labs and robotics companies have largely consumed the text available on the internet. The next gains come from data about people doing real tasks in real places: how a worker folds laundry, loads a dishwasher, packs an order or handles a customer return. That data does not exist on the web, and simulation does not reproduce it.
Realset’s answer is that every dataset starts with a real person doing a real task in a real place. The company uses expert demonstrators rather than crowd annotators, real environments rather than simulation, and delivers de-identified data to US-hosted cloud buckets, with quality measured at every step of a single pipeline.
Three Ways to Produce Ground Truth
Realset Body: embodied data for physical policy models. Egocentric and third-person capture of skilled workers performing manipulation tasks in homes, kitchens, warehouses and light assembly lines, plus bimanual teleoperation episodes with synchronized stereo video, IMU and action logs. Delivered with dense action-level annotation optimized for vision-language-action (VLA) training.Realset Field: LLM and agent training data from RL environments built on real workflows. Environments that mirror e-commerce operations, customer support, logistics dispatch and manufacturing SOPs. Domain experts generate trajectories, preference pairs and rubrics inside the environment, with verifiable rewards from real business outcomes and bilingual English/Chinese expert pools.Realset Judge: expert evaluation for agents in production. Evaluation design, failure diagnosis and continuous monitoring by people who do the job the agent is replacing, with targeted fix-data for the top failure modes and re-scoring as models and prompts change. The same services are offered to data integration and AI solution companies that deploy agents for their own clients.
The Realset Workspace
All three run on the Realset Workspace, where domain experts in six languages (English, Simplified Chinese, Japanese, Korean, Spanish and Arabic) answer real tasks, attach evidence and screen recordings, and pass an independent quality review before a record is approved. Every approved record exports as JSONL with provenance that customers can verify.
Open Benchmarks on Real Tasks
Realset publishes open benchmarks so the field can measure what matters: whether a policy works outside the lab. The Realset Household Manipulation Bench evaluates open-source VLA policies including π0, OpenVLA, GR00T and Octo on folding, loading, sorting and wiping tasks captured in real kitchens and laundry rooms, scored by success rate over three trials per task, with results expected in Q4 2026. A Light Assembly Bench, scored by the line workers who trained on the tasks, and a Commerce Ops Agent Bench for computer-use agents on real e-commerce seller operations are planned.
“The easy data is gone. What is left is the physical world, and you cannot scrape it,” said Hunter Guo, founder of Realset AI. “You have to put a camera on a skilled person doing real work, structure what they did, and check it with people who know the job. That is a capture and quality problem, not a labeling problem, and it is the problem we are building a company around.”
About Realset AI
Realset AI is a real-world data lab and training data provider for LLMs and embodied AI. It captures expert human demonstrations, builds RL environments from real workflows and evaluates AI agents with domain experts, for frontier labs, robotics companies, and data integration and AI solution providers. Realset is headquartered in San Jose, California. Learn more at https://realset.ai
About Flatkey
Flatkey is an AI infrastructure platform that gives developers access to more than 100 official AI models and more than 1,000 AI tools through one key and one balance. Learn more at https://flatkey.ai
Media Contact
Xingru Ren
Head of Marketing
+1 424 356 6176
xingru@flatkey.ai
https://realset.ai
View original content to download multimedia:https://www.prnewswire.com/news-releases/realset-ai-and-flatkey-raise-10m-series-a-to-build-real-world-training-data-for-frontier-models-and-embodied-agents-302887825.html
SOURCE Realset AI
