Real-time AI for household tasks
Humanfields is building an AI assistant that sees household tasks through your camera, understands what you are doing, and guides you through each step in real time, helping people handle the tasks they would normally outsource or avoid.
01
See the physical task through your camera.
02
Connect what it sees to the right instructions and procedures.
03
Guide each action while keeping you in control.
The premise
Robots are the destination. We are the bridge.
Full automation of physical tasks is where things are heading. Robots that assemble furniture, cook meals, and handle repairs without being asked. That technology is not ready for most homes yet.
The other problem is that most people do not know the right steps for everyday physical tasks. Assembly instructions are hard to follow. Recipes assume knowledge you do not have. Repairs look simple until you start.
Humanfields fills that gap. An AI that watches through your camera, understands the task you are working on, and tells you the next step in plain language. Faster, done right, by you.
Why now
Three things converged at once.
Vision models can now understand physical situations reliably. A model that could not identify furniture components two years ago can now do it from a phone camera.
The hardware is already in your pocket. A phone is enough to see the task, understand what is happening, and guide the next action. You do not need a new device.
The habits around physical tasks are shifting. People increasingly outsource things they could do themselves, not because the tasks are hard but because they do not know where to start. AI guidance changes that calculation.
The system
Watch, reason, and guide.
A useful assistant for physical tasks needs to understand the actual situation, not a typed description of it. It needs to see what you see, know what step you are on, and give you one clear instruction at a time.
Our prototype combines live visual understanding with structured task knowledge. It tracks your progress through the camera and adapts when things go differently from the plan.
Application domains
Starting at home.
The home is where the gap between wanting something done and knowing how to do it is most visible and most solvable.
Furniture assembly
Flat-pack, done right
You open the box, spread out 47 parts, and the diagram makes no sense. Our AI reads the manual, watches through your camera as you work, and walks you through each step in plain language. No restarts. No parts left over at the end.
↗ Complete builds faster
Design principles
Camera-first context for the actual task in front of you, not a generic version of it.
Real-time guidance that keeps pace with the task without taking over.
Guidance grounded in real manuals, procedures, and instructions.
Built to complete the task, not to extend the conversation.
Get in touch