cap-x vs PhyAgentOS-core

PhyAgentOS-core is much bigger: 2.1k stars against 769. Over the days we have tracked them cap-x moved +17.2% and PhyAgentOS-core +18.8%, so PhyAgentOS-core is growing faster right now.

They split the axes: cap-x leads on model freedom, PhyAgentOS-core on context depth and noise control and customization.

Stars and commit dates come from our own daily tracking. The six axes are read off each project's documentation by our review pipeline, so they describe what a project says about itself, not what we measured in its code.

Where they stand today

Provides an end-to-end, robotics-focused framework that benchmarks and improves Code-as-Policy agents with interactive robot gym environments, multi-tier benchmarks, a training-free agent, and RL fine-tuning tools bundled together.

Stars
769
Tracked growth
+17.2%
Maturity
Last commit
99d ago
Language
Python
License
MIT
Cost to run
Your API key or local GPU

Decouples cognition from physical hardware with a session-centered runtime and small target adapters so the same agent codebase runs across sim and real robots.

Stars
2.1k
Tracked growth
+18.8%
Maturity
Last commit
4h ago
Language
Python
License
MIT
Cost to run
Free, self-hosted (local compute)
0%+19%41 tracked days
capgym/cap-xPhyAgentOS/PhyAgentOS-core

Six axes, head to head

Each axis runs 0 to 5. The label under a score is what that project's own docs claim, not a category average.

Axiscap-xPhyAgentOS-core
Context depth
How much of your codebase it sees before it answers: the open diff, the diff plus related files, or the whole repository.
Diff only
Workspace-wide protocol
Noise control
How it keeps output volume down — severity thresholds, deduplication, incremental runs over new commits only.
None mentioned
Multi-layer safety
Customization
How far it bends to your team: custom rules, prompts, style guides, per-path config.
Config file
Plugins + YAML configs
Privacy
Whether your code stays on your own infrastructure: fully local, self-hostable, or cloud API only.
Self-hostable
Fully self-hosted
Model freedom
Whether you can point it at any provider, or it is wired to one.
Bring-your-own-key
Specific policy endpoints
Setup ease
What it takes to get a first useful run out of it.
Config + API key
Quick CLI onboarding

Which one to pick

Pick cap-x if…

Robotics-focused — pick CaP-X when you need a comprehensive, open-source benchmark and toolkit for evaluating and improving code-generating agents on robot manipulation tasks with simulation environments and RL support.

  • Model freedom: Bring-your-own-key (5/5 against 3/5)
Runs in cli, web-app, ci. Works with byok, gemini, other-fixed.

Pick PhyAgentOS-core if…

Safety-first — built-in multi-layer validation (critic → preflight → target-side SafetyGuard) and auditable file protocols make it suitable for real-robot deployment.

  • Context depth: Workspace-wide protocol (4/5 against 1/5)
  • Noise control: Multi-layer safety (5/5 against 1/5)
  • Customization: Plugins + YAML configs (5/5 against 3/5)
Runs in cli. Works with other-fixed.

What people want from each one

Questions people ask

Is cap-x better than PhyAgentOS-core?

They split the axes: cap-x leads on model freedom, PhyAgentOS-core on context depth and noise control and customization. cap-x is worth picking when robotics-focused — pick CaP-X when you need a comprehensive, open-source benchmark and toolkit for evaluating and improving code-generating agents on robot manipulation tasks with simulation environments and RL support.

Which of cap-x and PhyAgentOS-core keeps my code private?

cap-x: Self-hostable (4/5). PhyAgentOS-core: Fully self-hosted (5/5).

What does each one cost to run?

cap-x: Your API key or local GPU. PhyAgentOS-core: Free, self-hosted (local compute).

Full profiles: capgym/cap-x and PhyAgentOS/PhyAgentOS-core. Everything else in Robotics & embodied.