# robot-agent-exploration/rax

> An open-source harness where a general coding agent (Codex CLI or Claude Code) learns robot manipulation tasks through online exploration instead of demonstrations, then compiles what worked into a standalone modular program. Aimed at robotics and embodied-AI researchers.

- Magnitude: 3.0 out of 10 — Early signal
- Stars: 108 total · +9 stars measured 2026-10-19, 07:00–08:29 UTC, ≈ 26 by evening
- Star trust: star growth looks organic
- Category: AI agents · License: Apache-2.0 · Created: 2026-10-13 · Last push: 2026-10-18
- GitHub: https://github.com/robot-agent-exploration/rax · Page: https://gitnova.dev/en/r/robot-agent-exploration/rax

## Useful for

- Run RAX on Franka FR3 or Rainbow RB-Y1 hardware to train manipulation tasks without demos.
- Evaluate the compiled controller program on held-out simulation layouts in Docker.

## Why it’s here

- Star-counter measurements on 2026-10-19 (UTC), 07:00–08:29: 99 → 108 stars (+9). This is the change over that interval.
- Estimated end-of-day forecast: about +26 stars, using observed gains and the previous day.
- The repository is 6 days old and already has 108 stars. With less than two weeks of history, there's no usual pace to compare the spike against yet.
- Top new repositories this week: #178.
- Recent forks include notable developers: @mvanhorn (5,540 followers).

## Star trust

Star growth looks organic.

Star-trust labels are heuristics based on the repository’s behavior, not a check of every stargazer.

## Numbers

- Forks: 3
- Issues and pull requests: 3
- Watchers: 0
- Average over the last week: 18 per day
- Usual pace: too little history (under two weeks)
- Stars in the last hour (measured): 6

## Stars per day, last 7 days (oldest → newest, today is partial)

2026-10-13 … 2026-10-19: 22, 13, 14, 8, 15, 27, 9

## Spotted in now

- Top new repositories this week: #178

## Similar by description

1. **BinceQu/RoboHarness** — 0.0 · Cooling · AI agents · Python · Not enough measurements today · unusual star pattern (heuristic)
   A framework for robot control via an LLM agent: visual keypoints on a 2D image are tracked by optical flow and turned into geometric constraints for primitive actions, without training models.
   Full card: https://gitnova.dev/en/r/BinceQu/RoboHarness.md
2. **seohongpark/ocbench** — 0.7 · Steady · Science & research · Python · +1 star measured 2026-10-19, 00:14–08:39 UTC, ≈ 2 by evening
   OCBench is a controllable benchmark for robotic manipulation, designed for studying behavioral cloning (BC) and reinforcement learning (RL) algorithms. It provides GPU-accelerated tasks and scripted policies that mimic human…
   Full card: https://gitnova.dev/en/r/seohongpark/ocbench.md
3. **robosimstudio/robosimstudio** — 2.0 · Cooling · Hardware, IoT & robotics · Python · +5 stars measured 2026-10-19, 00:10–08:32 UTC, ≈ 13 by evening
   An open interactive GPU toolkit for rebuilding real robot manipulation tasks in simulation, by hand or with a coding agent (Real2Sim).
   Full card: https://gitnova.dev/en/r/robosimstudio/robosimstudio.md

---

Magnitude (0–10) measures how fast and how unusually interest in a repository is growing right now. It is not a quality score.
Days are UTC. “So far today” is a fact; “expected by the end of the day” is a forecast.
Summaries and use cases are written by an LLM (DeepSeek V4.1 Flash) from the README and may be inaccurate: verify specific claims (benchmarks, speed, hardware) in the repository itself.
Data as of 2026-10-19 08:40 UTC, updated every 30 minutes.