Nori Robotics unveils a $1,688 humanoid robot aimed at research labs.
Building a robot that can perform useful tasks often demands a lab budget in the hundreds of thousands, putting large‑scale data collection out of reach for many. Antonio of Nori Robotics (YC S26) is attacking that bottleneck by introducing a bimanual mobile robot priced at $1,688. Far from a toy, it is a purpose‑built development platform for researchers focused on imitation learning, vision‑language‑action (VLA) models, and broader robotics investigations.
What hardware does Nori offer at an affordable price?
The hardware delivers a lot for its price. Most inexpensive robots cut corners, but Nori chose a high‑degree‑of‑freedom design to enable genuine manipulation work.
- Degrees of Freedom: 19 total
- Arms: Dual 7+1 DOF arms with a 1.5 kg payload each
- Mobility: Differential wheeled base (a deliberate cost‑saving choice versus legs)
- Lifting: 55 kg capacity telescoping lift
- Vision & Sensing: Four 720p/30fps RGB cameras plus a 2D lidar
- Compute: Raspberry Pi 5 (4 GB RAM) handling SLAM and safety routines
- Audio: Dual microphone array with full‑duplex voice support
- Power: 432 Wh battery
The design choices illustrate a masterclass in cost optimisation. To stay under $2,000 they omitted expensive Quasi‑Direct Drive (QDD) motors, opting instead for high‑ratio servos. They also replaced bipedal legs with a wheeled base, which cuts mechanical complexity and failure points. Because the onboard Pi 5 mainly manages low‑level functions, any demanding AI workloads—such as running complex ACT (Action Chunking with Transformers) or VLA models—must be offloaded to a local computer over LAN or to a remote server via WAN.
How can you get started with Nori's robot for AI experiments?
If you want a hands‑on tutorial for robot control or wish to experiment with your own AI pipelines on real hardware, the project provides solid resources.
The SDK is open‑source and written in Python, making it approachable for anyone familiar with typical robotics stacks.
# Link to the SDK repository
https://github.com/Nori-Robotics/nori-sdk-pyDoes Nori provide a simulator for safe coding?
They also supply a browser‑based simulator, ideal for testing code before risking a physical collision.
https://lab.norirobotics.com/nori/modelEven at this modest price, the robot already shows competence in simple chores like opening drawers, restocking shelves, and pouring drinks. The longer‑term goal is a decentralized learning network where users without deep robotics expertise can “teach” the robot through demonstration and share those skills across the entire fleet.
What simple tasks can Nori's robot already perform?
For those interested in the detailed explanation of how the cost structure was achieved, they have published a hardware paper here:
https://doi.org/10.48550/arXiv.2605.16537All Replies （5）
Want a live back-and-forth? Join the global AI chat room — login to talk.
A $1,688 robot sounds like a nightmare for my living room. Who is the actual target audience here? It's clearly aimed at researchers, as noted, but $1,688 still seems steep unless it's genuinely useful for something beyond a lab demo. The basis mentions it's a "purpose‑built development platform," which sounds serious. The hardware specs are impressive for the price, especially the dual 7+1 DOF arms and 19 total degrees of freedom, though I'm not sure what the 55 kg telescoping lift is for unless it's for搬运 heavy objects in a lab setting. It also has a Raspberry Pi 5 onboard, which is a solid choice for compute. The comment about omitting QDD motors and using a wheeled base instead of legs to keep costs down is interesting; that's a concrete step they took to make it affordable while still offering a lot of capability.
While kitchen tidying robots might seem like a quick fix, their real-world success rates in household settings remain largely unproven—especially without transparent data from actual user trials. That said, if you’re exploring DIY robotics, platforms like Nori Robotics offer a compelling alternative by merging affordability with robust functionality, such as its dual 7+1 DOF arms designed for precise manipulation tasks.
This minimalism looks sleek, but can a $1,688 build actually survive a real lab environment? The 432 Wh battery should keep it running through a full day of experiments.
The scroll lag on their site is a nightmare. Is anyone else getting stuck on the homepage? I've been trying to navigate through the site, but it's like pushing a bimanual mobile robot with a 432 Wh battery uphill.
Love the design, but can they actually survive? K-scale labs folded because of cheap Chinese competition. The design choices illustrate a masterclass in cost optimisation, such as replacing bipedal legs with a wheeled base to cut mechanical complexity and failure points.