Dream Labs details its robot-control API and Dream-1

B Capital led Dream Labs’ funding round, with Kindred Ventures, A.Capital, Unusual Ventures, SV Angel, Wing VC, Scribble VC and Adrian Aoun among the participants.[2] Dream Labs says it was founded in 2026 and is headquartered in San Francisco.[2] The company says its founders are former NVIDIA researchers who worked on DreamZero and World Action Models.[2] Dream Labs says it will use the funding primarily for compute as it expands Dream-1’s capabilities and capacity.[4] The evidence does not specify the round’s total amount.

Dream Labs says Dream-1 learns from videos of people performing everyday tasks and predicts both what happens next and the robot’s resulting action.[2] The company says the model requires no fine-tuning or demonstrations from the robot, and describes its zero-shot scope as bimanual robots with parallel grippers.[3] Dream Labs says Dream-1 can control new two-armed robots with parallel grippers out of the box.[3] Its current tasks include tabletop picking and placing, as well as folding laundry.[2] Dream Labs plans to add support for wheeled mobile robots and dexterous hands.[2]

Dream API converts visual inputs and natural-language prompts into robot actions.[2] Dream Labs describes it as the first publicly available API that turns a robot’s camera feed and a plain-English instruction into motor commands.[2] Developers and robot makers can submit demonstration data and receive a tailored model within hours.[2] The company also says its Robot Playground lets anyone with a browser send commands to a robot in its San Francisco office and watch the response live, without buying a robot.[2] Dream Labs says it has more than 20 hardware partners and supports robots costing about $2,000 to $10,000.[2]

Dream Labs lists Dream-1 at $3 per hour and Dream-1-flash at $2 per hour.[3] Dream Labs says each model call returns 1.6 seconds of actions, or 2,250 calls per hour.[3] Dream-1 uses three camera views: one scene view and one view on each wrist.[4] For teams assessing a deployment, the listed robot scope, task fit and pricing tiers are important constraints to check before estimating operating costs.