Instance
ActiveAI-nativeAutomated evals for robot policies
What it does
We’re automating robot evaluation– the loop where you wait for your robot to roll out, judge its success, and then reset the scene. We're starting with step one: a success detector that automatically tells you whether your robot actually completed its task along with detailed subtask breakdown. We're MIT computer scientists & friends since middle school, with backgrounds at NASA JPL, SpaceX, and AWS. Book a demo with founders@instancelabs.ai.
Its site says now · captured 2026-08-23
Instance — Ground truth for robot learning
Instance verifies whether robots actually succeed at their tasks — reliable success labels for training, evaluation, and reinforcement learning.
Next to it in Agent infrastructure · Robot policy evaluation automation
- RindlerY Combinator S26
The translation layer between AI agents and the web
- HoneyHivePlug and Play PnP 2026
Modern AI Observability and Evaluation. Your single platform to observe, evaluate, and improve AI agents — whether you're just getting started or scaling agents across your enterprise.
- MinicorY Combinator X26
Scalable windows desktop automations
- StableBrowseY Combinator X26
Help frontier labs build better models.
- FernY Combinator W26
RL environments for robotics companies
- Ressl AIY Combinator W26
Train, eval and build autonomous agents
- RadiusY Combinator F25
Code and browse in the same place
- KERNELY Combinator S25
Crazy fast, open source infra for AI agents to use the Internet