{
  "schema_version": "openagent.resource.v1",
  "id": "res_rlinf",
  "slug": "rlinf",
  "status": "published",
  "identity": {
    "name": "RLinf",
    "one_liner": "Production-grade reinforcement learning infrastructure for embodied and agentic AI.",
    "short_description": "RLinf is a flexible and scalable open-source RL infrastructure designed for Embodied and Agentic AI. It supports real-world robot RL on Franka, XSquare Turtle2, and DOS-W1 arms, multiple simulation backends (ManiSkill, LIBERO, MetaWorld, IsaacLab, RoboCasa), and state-of-the-art VLA model fine-tuning (Pi0, Pi0.5, GR00T, OpenVLA). It also extends to agentic AI with support for Search-R1, rStar2, and multi-agent RL."
  },
  "classification": {
    "resource_type": "bot",
    "primary_category": "bots",
    "subcategories": [
      "robotics",
      "reinforcement-learning",
      "vla",
      "python",
      "gpu"
    ]
  },
  "positioning": {
    "why_it_matters": "RLinf matters because reinforcement learning for robotics has been fragmented across incompatible tools, simulators, and algorithms. RLinf provides a unified infrastructure that works across simulation, real-world robots, and even agentic AI — reducing the engineering overhead of setting up RL experiments from weeks to hours. Its support for real-world online RL (HG-DAgger) and production-grade RL algorithms (PPO, GRPO, SAC, DAPO) makes it one of the most comprehensive open RL frameworks available.",
    "best_for": [
      "Robotics researchers running RL experiments across simulation and real hardware",
      "Teams fine-tuning VLA models with reinforcement learning",
      "Developers building agentic AI systems with RL-based training"
    ],
    "not_for": [
      "Beginners looking for a simple out-of-the-box robot control interface (start with LeRobot)"
    ],
    "use_cases": [
      "robotics-agent"
    ],
    "target_audience": [
      "developer",
      "agent_builder"
    ],
    "maturity": "active"
  },
  "decision_signals": {
    "deployment_modes": [
      "cloud"
    ],
    "open_source": true,
    "local_first": false,
    "self_hostable": false,
    "has_api": false,
    "has_gui": false,
    "supports_mcp": false,
    "supports_docker": false
  },
  "facts": {
    "license": "Apache-2.0",
    "pricing_model": "open_source",
    "github_stars": 3161,
    "github_forks": 411,
    "github_repo_full_name": "RLinf/RLinf",
    "last_verified_at": "2026-06-04"
  },
  "capabilities": {
    "core_capabilities": [
      "robotics",
      "messaging"
    ],
    "interfaces": [
      "repo",
      "docs"
    ]
  },
  "links": {
    "primary_url": "https://github.com/RLinf/RLinf",
    "items": [
      {
        "type": "github",
        "label": "GitHub",
        "url": "https://github.com/RLinf/RLinf"
      },
      {
        "type": "homepage",
        "label": "Homepage",
        "url": "https://rlinf.readthedocs.io/en/latest/"
      }
    ]
  },
  "media": {
    "thumbnail_brief": {
      "resource_type": "bot",
      "visual_motif": "RL training loop diagram with robot arm and neural network",
      "background_style": "quiet editorial card with light surface and orange accent",
      "title_overlay": "RLinf",
      "subtitle": "RL infrastructure for embodied AI",
      "avoid": [
        "dense algorithm pseudocode",
        "dark control panel look"
      ]
    }
  },
  "tags": {
    "category": [
      "bot",
      "open-source"
    ],
    "capability": [
      "robotics",
      "messaging"
    ],
    "constraint": [
      "open-source"
    ],
    "scenario": [
      "robotics-agent"
    ]
  },
  "relationships": {},
  "machine_readable": {
    "canonical_url": "https://www.openagent.bot/bots/rlinf",
    "json_url": "https://www.openagent.bot/bots/rlinf.json",
    "markdown_url": "https://www.openagent.bot/bots/rlinf.md"
  },
  "seo": {
    "title": "RLinf: Open-source RL infrastructure for embodied and agentic AI",
    "description": "RLinf is production-grade open-source reinforcement learning infrastructure for robotics and agentic AI. Supports real-world RL, VLA fine-tuning, and multi-agent systems."
  },
  "editorial": {
    "trust_note": "Verified from source links and project metadata.",
    "core_strengths": [
      {
        "title": "Unified RL across simulation and real hardware",
        "description": "RLinf supports 10+ simulation backends (ManiSkill, LIBERO, MetaWorld, IsaacLab, RoboCasa, Calvin, etc.) and real-world robots (Franka, XSquare Turtle2, DOS-W1) with the same API.",
        "why_it_matters": "You can prototype in simulation and deploy on real hardware without rewriting your RL pipeline."
      },
      {
        "title": "State-of-the-art VLA RL fine-tuning",
        "description": "Fine-tune Pi0, Pi0.5, GR00T, OpenVLA, LingBot-VLA and other VLA models using RL algorithms like GRPO, PPO, and DAPO.",
        "why_it_matters": "VLA models are typically trained with imitation learning only. RLinf enables RL-based post-training that can surpass demonstration quality."
      },
      {
        "title": "Real-world online RL with HG-DAgger",
        "description": "Human-Gated DAgger allows safe online RL on real robots — a human supervisor gates when the policy's actions are used vs. when human corrections are needed.",
        "why_it_matters": "Online RL on real hardware is dangerous without safety mechanisms. HG-DAgger provides a practical bridge between human demonstrations and autonomous RL."
      },
      {
        "title": "Agentic AI support",
        "description": "Extends beyond robotics to support RL for language agents — Search-R1, rStar2, coding agents, and multi-agent systems.",
        "why_it_matters": "RLinf is one of the few frameworks that bridges embodied RL and agentic RL in a single codebase."
      }
    ],
    "use_case_notes": [
      {
        "title": "RL-based post-training for VLA policies",
        "description": "After collecting demonstration data and training a VLA policy with imitation learning, use RLinf to fine-tune the policy with RL for higher success rates."
      },
      {
        "title": "Real-world robot learning with safety guarantees",
        "description": "Deploy RLinf on a Franka arm with HG-DAgger for safe online learning — the human intervenes when the policy makes unsafe moves, and the system learns from both successes and corrections."
      },
      {
        "title": "Multi-agent embodied RL research",
        "description": "Use RLinf's multi-agent support to study coordination between multiple robots performing collaborative tasks in simulation."
      }
    ],
    "compare_notes": [
      {
        "title": "Choose RLinf for production RL across robots and agents",
        "summary": "Stable-Baselines3 is simpler for standard RL benchmarks but lacks robot integration. RLinf provides the full stack from simulation to real hardware to agentic AI.",
        "against": "specialized RL libraries"
      }
    ],
    "getting_started": [
      {
        "label": "View the GitHub repository",
        "url": "https://github.com/RLinf/RLinf",
        "type": "github"
      },
      {
        "label": "Read the documentation",
        "url": "https://rlinf.readthedocs.io/en/latest/",
        "type": "docs"
      }
    ],
    "command_line": [
      {
        "label": "Install RLinf",
        "command": "pip install rlinf",
        "description": "Install RLinf from PyPI."
      }
    ],
    "seo_article": {
      "intro": "RLinf is a production-grade open-source reinforcement learning infrastructure that unifies embodied AI robotics and agentic AI language models under one RL framework.",
      "what_it_is": "RLinf is a flexible and scalable RL infrastructure supporting 10+ simulation backends, real-world robot control, VLA model fine-tuning, and agentic AI. It implements major RL algorithms (PPO, GRPO, SAC, DAPO, IQL, CrossQ, RLPD) with a unified API that works identically across simulation and real hardware. Its real-world RL stack includes HG-DAgger for safe online training, and its agentic AI module extends RL to language agents.",
      "why_it_matters": "Reinforcement learning for embodied AI has been held back by the gap between simulation research and real-world deployment. RLinf bridges this gap by providing the same API across 10+ simulators and multiple real robot platforms. It also bridges the gap between robotics RL and agentic RL — a convergence that is increasingly important as VLA models and language agents share architectures and training techniques.",
      "how_it_works": "RLinf provides a modular architecture where environments, policies, and algorithms are swappable components. An experiment is configured via YAML or Python dict, specifying the simulator backend (or real robot), the policy model (from MLP to VLA), and the RL algorithm. For real-world RL, the HG-DAgger loop runs a policy on hardware, a human supervisor monitors and intervenes via a GUI, and the system logs both autonomous and human-corrected episodes for training.",
      "use_cases": [
        {
          "title": "Fine-tuning Pi0 with RL for higher manipulation success",
          "description": "Train Pi0 with imitation learning on demonstrations, then run RL post-training with RLinf to improve success rates beyond the demonstrator's performance."
        },
        {
          "title": "Real-world Franka arm learning with HG-DAgger",
          "description": "Set up a Franka arm with ZED cameras and Robotiq gripper, run online RL with human-gated intervention for safe exploration on real pick-and-place tasks."
        },
        {
          "title": "Agentic RL for search and reasoning",
          "description": "Use RLinf's Search-R1 and rStar2 support to apply RL training to language agent search and reasoning tasks, improving performance beyond supervised fine-tuning."
        }
      ],
      "alternatives": [
        {
          "title": "Use Stable-Baselines3 for simpler benchmark RL",
          "summary": "SB3 is more lightweight for standard Gymnasium benchmarks. RLinf is the choice when you need robot hardware integration, VLA support, or multi-agent coordination.",
          "against": "RLinf"
        }
      ],
      "getting_started": [
        {
          "label": "Clone the repository",
          "url": "https://github.com/RLinf/RLinf",
          "type": "github"
        },
        {
          "label": "Read the docs",
          "url": "https://rlinf.readthedocs.io/en/latest/",
          "type": "docs"
        }
      ],
      "faq": [
        {
          "question": "What RL algorithms does RLinf support?",
          "answer": "RLinf supports IQL, GRPO, PPO, DAPO, Reinforce++, SAC, CrossQ, RLPD, SAC-Flow, DSRL, and RECAP/CFG among others."
        },
        {
          "question": "What robots are supported for real-world RL?",
          "answer": "Franka Arm (with RealSense, ZED cameras, Franka Hand, Robotiq gripper), XSquare Turtle2 dual-arm, and DOS-W1. More robots are being added."
        },
        {
          "question": "Can I use RLinf without real hardware?",
          "answer": "Yes, RLinf supports 10+ simulation backends including ManiSkill, LIBERO, MetaWorld, IsaacLab, RoboCasa, Calvin, and more — all accessible with the same API."
        }
      ]
    }
  },
  "timestamps": {
    "created_at": "2026-06-04T00:00:00.000Z",
    "updated_at": "2026-06-04T00:00:00.000Z",
    "published_at": "2026-06-04T00:00:00.000Z"
  }
}