VectorRobotics

VectorRobotics / vector-os-nano

Public

Vector OS Nano: Natural language controlled robot arm. $450 hardware, say 'pick up the battery' and it does. LeRobot SO-ARM100 + RealSense D405 + ROS2 + LLM.

16
0
100% credibility
Found Mar 21, 2026 at 16 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
Python
AI Summary

Vector OS Nano is an open-source ROS2 software stack that enables a low-cost ($420) robot arm with depth camera to perform zero-shot grasping of arbitrary objects via natural language commands.

How It Works

1
🔍 Discover the magic

You watch a fun video of a tiny robot arm picking up everyday objects just by hearing simple commands like 'pick the battery'.

2
🛒 Gather your robot parts

Order a cheap robot arm and camera for under $500, plus use your existing computer with a graphics card.

3
💻 Prepare your setup

Follow easy steps to get the free software ready on your Ubuntu computer.

4
📐 Teach your robot its space

Spend 5-10 minutes placing objects in different spots so the robot learns where to reach.

5
🧠 Connect the smart brain

Link a helpful AI service so your robot understands everyday words in English or Chinese.

6
▶️ Wake up your robot

Open two simple windows and start the system with one command each.

7
🗣️ Talk to your robot

Type natural sentences like 'grab the red cup' or 'pick up the battery' and watch it go!

✅ Perfect picks every time

Your robot confidently grasps any object you describe, no training needed – ready for fun experiments!

Sign up to see the full architecture

6 more

Sign Up Free

Star Growth

See how this repo grew from 16 to 16 stars Sign Up Free
Repurpose This Repo

Repurpose is a Pro feature

Generate ready-to-use prompts for X threads, LinkedIn posts, blog posts, YouTube scripts, and more -- with full repo context baked in.

Unlock Repurpose
AI-Generated Review

What is vector-os-nano?

Vector OS Nano lets you command a $450 robot arm with natural language: type "pick up the battery" or "捡起桌上的电池" in a ROS2 CLI, and it detects via RealSense D405, tracks at 20fps, computes IK, and grasps zero-shot—no training needed. Python-based, it integrates LeRobot SO-ARM100 hardware, local VLM for scene description, and Claude Haiku LLM for bilingual intent parsing into pick(), home(), or detect() actions. Users get a full perception-to-action pipeline on Ubuntu with NVIDIA GPU.

Why is it gaining traction?

Affordable $450 setup (arm plus camera) delivers real grasping without sims or datasets, unlike rigid github vector robot scripts. CLI handles fuzzy queries like "grab the red cup" via LLM tool-calling, with auto-retry and gripper compensation—devs notice reliable 640x480@30fps picks fast. CMU Robotics Institute backing promises SLAM/navigation expansions.

Who should use this?

ROS2 robotics hobbyists building manipulation prototypes, researchers exploring LLM control on cheap arms, or github vector dev teams needing nano vector analyzer precision for pick/place tasks. Suits arm experimenters tired of MoveIt tuning or manual IK.

Verdict

Early-stage gem (16 stars, 1.0% credibility) with excellent quickstart/docs but light tests—calibrate workspace first. Try for $450 LLM-robotics if you're prototyping; star for full Vector OS releases.

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.