Metadata-Version: 2.4
Name: positronic
Version: 0.1.0
Summary: Positronic robotics project
License:                                  Apache License
                                   Version 2.0, January 2004
                                http://www.apache.org/licenses/
        
           TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION
        
           1. Definitions.
        
              "License" shall mean the terms and conditions for use, reproduction,
              and distribution as defined by Sections 1 through 9 of this document.
        
              "Licensor" shall mean the copyright owner or entity authorized by
              the copyright owner that is granting the License.
        
              "Legal Entity" shall mean the union of the acting entity and all
              other entities that control, are controlled by, or are under common
              control with that entity. For the purposes of this definition,
              "control" means (i) the power, direct or indirect, to cause the
              direction or management of such entity, whether by contract or
              otherwise, or (ii) ownership of fifty percent (50%) or more of the
              outstanding shares, or (iii) beneficial ownership of such entity.
        
              "You" (or "Your") shall mean an individual or Legal Entity
              exercising permissions granted by this License.
        
              "Source" form shall mean the preferred form for making modifications,
              including but not limited to software source code, documentation
              source, and configuration files.
        
              "Object" form shall mean any form resulting from mechanical
              transformation or translation of a Source form, including but
              not limited to compiled object code, generated documentation,
              and conversions to other media types.
        
              "Work" shall mean the work of authorship, whether in Source or
              Object form, made available under the License, as indicated by a
              copyright notice that is included in or attached to the work
              (an example is provided in the Appendix below).
        
              "Derivative Works" shall mean any work, whether in Source or Object
              form, that is based on (or derived from) the Work and for which the
              editorial revisions, annotations, elaborations, or other modifications
              represent, as a whole, an original work of authorship. For the purposes
              of this License, Derivative Works shall not include works that remain
              separable from, or merely link (or bind by name) to the interfaces of,
              the Work and Derivative Works thereof.
        
              "Contribution" shall mean any work of authorship, including
              the original version of the Work and any modifications or additions
              to that Work or Derivative Works thereof, that is intentionally
              submitted to Licensor for inclusion in the Work by the copyright owner
              or by an individual or Legal Entity authorized to submit on behalf of
              the copyright owner. For the purposes of this definition, "submitted"
              means any form of electronic, verbal, or written communication sent
              to the Licensor or its representatives, including but not limited to
              communication on electronic mailing lists, source code control systems,
              and issue tracking systems that are managed by, or on behalf of, the
              Licensor for the purpose of discussing and improving the Work, but
              excluding communication that is conspicuously marked or otherwise
              designated in writing by the copyright owner as "Not a Contribution."
        
              "Contributor" shall mean Licensor and any individual or Legal Entity
              on behalf of whom a Contribution has been received by Licensor and
              subsequently incorporated within the Work.
        
           2. Grant of Copyright License. Subject to the terms and conditions of
              this License, each Contributor hereby grants to You a perpetual,
              worldwide, non-exclusive, no-charge, royalty-free, irrevocable
              copyright license to reproduce, prepare Derivative Works of,
              publicly display, publicly perform, sublicense, and distribute the
              Work and such Derivative Works in Source or Object form.
        
           3. Grant of Patent License. Subject to the terms and conditions of
              this License, each Contributor hereby grants to You a perpetual,
              worldwide, non-exclusive, no-charge, royalty-free, irrevocable
              (except as stated in this section) patent license to make, have made,
              use, offer to sell, sell, import, and otherwise transfer the Work,
              where such license applies only to those patent claims licensable
              by such Contributor that are necessarily infringed by their
              Contribution(s) alone or by combination of their Contribution(s)
              with the Work to which such Contribution(s) was submitted. If You
              institute patent litigation against any entity (including a
              cross-claim or counterclaim in a lawsuit) alleging that the Work
              or a Contribution incorporated within the Work constitutes direct
              or contributory patent infringement, then any patent licenses
              granted to You under this License for that Work shall terminate
              as of the date such litigation is filed.
        
           4. Redistribution. You may reproduce and distribute copies of the
              Work or Derivative Works thereof in any medium, with or without
              modifications, and in Source or Object form, provided that You
              meet the following conditions:
        
              (a) You must give any other recipients of the Work or
                  Derivative Works a copy of this License; and
        
              (b) You must cause any modified files to carry prominent notices
                  stating that You changed the files; and
        
              (c) You must retain, in the Source form of any Derivative Works
                  that You distribute, all copyright, patent, trademark, and
                  attribution notices from the Source form of the Work,
                  excluding those notices that do not pertain to any part of
                  the Derivative Works; and
        
              (d) If the Work includes a "NOTICE" text file as part of its
                  distribution, then any Derivative Works that You distribute must
                  include a readable copy of the attribution notices contained
                  within such NOTICE file, excluding those notices that do not
                  pertain to any part of the Derivative Works, in at least one
                  of the following places: within a NOTICE text file distributed
                  as part of the Derivative Works; within the Source form or
                  documentation, if provided along with the Derivative Works; or,
                  within a display generated by the Derivative Works, if and
                  wherever such third-party notices normally appear. The contents
                  of the NOTICE file are for informational purposes only and
                  do not modify the License. You may add Your own attribution
                  notices within Derivative Works that You distribute, alongside
                  or as an addendum to the NOTICE text from the Work, provided
                  that such additional attribution notices cannot be construed
                  as modifying the License.
        
              You may add Your own copyright statement to Your modifications and
              may provide additional or different license terms and conditions
              for use, reproduction, or distribution of Your modifications, or
              for any such Derivative Works as a whole, provided Your use,
              reproduction, and distribution of the Work otherwise complies with
              the conditions stated in this License.
        
           5. Submission of Contributions. Unless You explicitly state otherwise,
              any Contribution intentionally submitted for inclusion in the Work
              by You to the Licensor shall be under the terms and conditions of
              this License, without any additional terms or conditions.
              Notwithstanding the above, nothing herein shall supersede or modify
              the terms of any separate license agreement you may have executed
              with Licensor regarding such Contributions.
        
           6. Trademarks. This License does not grant permission to use the trade
              names, trademarks, service marks, or product names of the Licensor,
              except as required for reasonable and customary use in describing the
              origin of the Work and reproducing the content of the NOTICE file.
        
           7. Disclaimer of Warranty. Unless required by applicable law or
              agreed to in writing, Licensor provides the Work (and each
              Contributor provides its Contributions) on an "AS IS" BASIS,
              WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
              implied, including, without limitation, any warranties or conditions
              of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A
              PARTICULAR PURPOSE. You are solely responsible for determining the
              appropriateness of using or redistributing the Work and assume any
              risks associated with Your exercise of permissions under this License.
        
           8. Limitation of Liability. In no event and under no legal theory,
              whether in tort (including negligence), contract, or otherwise,
              unless required by applicable law (such as deliberate and grossly
              negligent acts) or agreed to in writing, shall any Contributor be
              liable to You for damages, including any direct, indirect, special,
              incidental, or consequential damages of any character arising as a
              result of this License or out of the use or inability to use the
              Work (including but not limited to damages for loss of goodwill,
              work stoppage, computer failure or malfunction, or any and all
              other commercial damages or losses), even if such Contributor
              has been advised of the possibility of such damages.
        
           9. Accepting Warranty or Additional Liability. While redistributing
              the Work or Derivative Works thereof, You may choose to offer,
              and charge a fee for, acceptance of support, warranty, indemnity,
              or other liability obligations and/or rights consistent with this
              License. However, in accepting such obligations, You may act only
              on Your own behalf and on Your sole responsibility, not on behalf
              of any other Contributor, and only if You agree to indemnify,
              defend, and hold each Contributor harmless for any liability
              incurred by, or claims asserted against, such Contributor by reason
              of your accepting any such warranty or additional liability.
        
           END OF TERMS AND CONDITIONS
        
           APPENDIX: How to apply the Apache License to your work.
        
              To apply the Apache License to your work, attach the following
              boilerplate notice, with the fields enclosed by brackets "[]"
              replaced with your own identifying information. (Don't include
              the brackets!)  The text should be enclosed in the appropriate
              comment syntax for the file format. We also recommend that a
              file or class name and description of purpose be included on the
              same "printed page" as the copyright notice for easier
              identification within third-party archives.
        
           Copyright 2025 Positronic Robotics Inc.
        
           Licensed under the Apache License, Version 2.0 (the "License");
           you may not use this file except in compliance with the License.
           You may obtain a copy of the License at
        
               http://www.apache.org/licenses/LICENSE-2.0
        
           Unless required by applicable law or agreed to in writing, software
           distributed under the License is distributed on an "AS IS" BASIS,
           WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
           See the License for the specific language governing permissions and
           limitations under the License.
        
        
Classifier: License :: OSI Approved :: Apache Software License
Classifier: Programming Language :: Python :: 3
Classifier: Programming Language :: Python :: 3 :: Only
Requires-Python: >=3.11
Description-Content-Type: text/markdown
License-File: LICENSE
License-File: NOTICE
Requires-Dist: av
Requires-Dist: configuronic>=0.2
Requires-Dist: cvxopt
Requires-Dist: dearpygui
Requires-Dist: dm_control
Requires-Dist: fastapi
Requires-Dist: fire
Requires-Dist: imageio[ffmpeg]
Requires-Dist: mujoco
Requires-Dist: numpy
Requires-Dist: opencv-python-headless
Requires-Dist: pyarrow
Requires-Dist: pyturbojpeg
Requires-Dist: robosuite
Requires-Dist: numba>=0.57
Requires-Dist: ruckig
Requires-Dist: scipy
Requires-Dist: torch
Requires-Dist: uvicorn[standard]
Requires-Dist: lerobot>=0.3.3
Requires-Dist: rerun-sdk>=0.22
Provides-Extra: hardware
Requires-Dist: feetech-servo-sdk; extra == "hardware"
Requires-Dist: positronic-franka; sys_platform == "linux" and extra == "hardware"
Requires-Dist: kortex_api; sys_platform == "linux" and extra == "hardware"
Requires-Dist: linuxpy; extra == "hardware"
Requires-Dist: pinocchio; extra == "hardware"
Requires-Dist: placo; extra == "hardware"
Requires-Dist: pyaudio; sys_platform == "linux" and extra == "hardware"
Requires-Dist: pymodbus; extra == "hardware"
Requires-Dist: pyserial; extra == "hardware"
Requires-Dist: pyzed; sys_platform == "linux" and extra == "hardware"
Provides-Extra: dev
Requires-Dist: flake8; extra == "dev"
Requires-Dist: ipython; extra == "dev"
Requires-Dist: jupyter; extra == "dev"
Requires-Dist: opencv-python; extra == "dev"
Requires-Dist: coverage; extra == "dev"
Requires-Dist: pytest-cov; extra == "dev"
Requires-Dist: pytest-asyncio; extra == "dev"
Requires-Dist: pytest; extra == "dev"
Requires-Dist: rerun-sdk[notebook]; extra == "dev"
Dynamic: license-file

# Positronic — Python-native stack for real-life ML robotics

[![Python](https://img.shields.io/badge/Python-3.11+-blue.svg)](https://www.python.org/downloads/)
[![License](https://img.shields.io/badge/License-Apache_2.0-blue.svg)](LICENSE)
[![Tests](https://github.com/Positronic-Robotics/positronic/actions/workflows/unit-test.yaml/badge.svg)](https://github.com/Positronic-Robotics/positronic/actions/workflows/unit-test.yaml)
[![PyPI](https://img.shields.io/pypi/v/positronic)](https://pypi.org/project/positronic/)

## What is Positronic

Positronic is an end-to-end toolkit for building ML-driven robotics systems.

It covers the full lifecycle:
- bring new hardware online,
- capture and curate datasets,
- train and evaluate policies,
- deploy inference,
- monitor performance,
- and iterate when behaviour drifts.

Every subsystem is implemented in plain Python so teams can move between simulation, teleoperation, and production code without switching languages or tools.

**Our goal is to make professional-grade ML robotics approachable.** Newcomers get clear defaults, WebXR teleoperation, and a reproducible training story; experienced teams can extend the same stack with custom drivers, schedulers, and datasets to support complex manipulation setups and large fleets.

> **Positronic is under heavy development and in alpha stage. APIs, interfaces, and workflows may change significantly between releases.**


## Table of Contents
- [Why Positronic](#why-positronic)
- [Installation](#installation)
- [Core modules at a glance](#core-modules-at-a-glance)
- [End-to-end workflow](#end-to-end-workflow)
  - [1. Prepare the environment](#1-prepare-the-environment)
  - [2. Collect demonstrations](#2-collect-demonstrations)
  - [3. Review and curate datasets](#3-review-and-curate-datasets)
  - [4. Convert to LeRobot](#4-convert-to-lerobot)
  - [5. Train a policy](#5-train-a-policy)
  - [6. Validate policies](#6-validate-policies)
- [Tooling and further reading](#tooling-and-further-reading)
- [Development workflow](#development-workflow)
- [How Positronic differs from LeRobot](#how-positronic-differs-from-lerobot)
- [Roadmap](#roadmap)

## Why Positronic

- **Immediate-mode runtime.** [`pimm`](pimm/README.md) wires sensors, controllers, inference, and GUIs without ROS launch files or bespoke DSLs. Loops stay testable and readable; see the [Pimm README](pimm/README.md) for details.
- **Dataset infrastructure.** The [Positronic dataset library](positronic/dataset/README.md) layer captures multi-rate episodes, exposes lazy transforms, and feeds the web-based browser for triage and analysis (see the [Dataset README](positronic/dataset/README.md)).
- **Hardware-ready out of the box.** [WebXR driver](positronic/drivers/webxr.py), [MuJoCo integration](positronic/simulator/mujoco/sim.py), the [DearPyGui UI](positronic/gui/dpg.py), and the [Drivers package](positronic/drivers) live alongside training and inference scripts so you can go from prototype to lab deployment quickly.
- **Future-facing roadmap.** Native training on Positronic datasets, curation capabilities in the viewer server, out-of-the-box support for $\pi_0$, GR00T and other polcies, support for more robotic and controlling hardware are active areas of development. Everything in this README reflects the workflow we run today and the direction we are pushing toward.

---

## Installation

Clone the repository and set up a local `uv` environment (recommended) or use Docker if you prefer containers.

### Option 1: Local via uv (recommended)

Prerequisites: Python 3.11 and [uv](https://docs.astral.sh/uv/).

```bash
git clone git@github.com:Positronic-Robotics/positronic.git
cd positronic

uv venv -p 3.11               # optional but keeps the interpreter isolated
source .venv/bin/activate     # activate the venv if you created one
uv sync --frozen --extra dev  # install core + dev tooling
```

Install hardware extras only when you need physical robot drivers:

```bash
uv sync --frozen --extra hardware
```

All runtime commands in the sections below assume either an activated virtual environment or `uv run --with-editable . -s …`.

### Option 2: Docker

```bash
git clone git@github.com:Positronic-Robotics/positronic.git
cd positronic

./docker/build.sh            # build image (includes libfranka and uv.lock deps)
./docker/run.sh pytest       # run commands inside the container
```
---

## Core modules at a glance

- [`pimm`](pimm/README.md) — immediate-mode runtime for building control systems. Handy references: [README](pimm/README.md), [Data collection example](positronic/data_collection.py), [Inference example](positronic/run_inference.py).
- [Positronic dataset library](positronic/dataset/README.md) — dataset writer/reader, transforms, and streaming agent.
- [Positronic server](positronic/server/positronic_server.py) — FastAPI + Rerun viewer for inspecting recordings.
- [Training scripts](positronic/training) — scripts for converting datasets and running LeRobot pipelines while native Positronic training is being finalised.
- [Drivers package](positronic/drivers) — hardware definitions, WebXR frontends, simulator loaders, and [Config presets](positronic/cfg) ready to override per project.

---

## End-to-end workflow

The usual loop is: collect demonstrations → inspect and curate → (temporarily) convert to LeRobot → train → validate. The sections below link straight to the scripts that implement each step.

### 1. Prepare the environment

- Create or activate your Python environment as shown above.
- For simulation-only workflows no additional extras are required.
- Hardware users should also provision device-level dependencies (e.g., udev rules for cameras) as described in the relevant driver documentation under [Drivers package](positronic/drivers/).

### 2. Collect demonstrations

Use the [Data collection script](positronic/data_collection.py) for both simulation and hardware captures.

#### Quick start in simulation

```bash
python -m positronic.data_collection sim \
    --output_dir=~/datasets/stack_cubes_raw \
    --webxr=.oculus \
    --operator_position=OperatorPosition.BACK
```

This command loads the [MuJoCo scene](positronic/assets/mujoco/franka_table.xml) with loaders from the [simulator config](positronic/cfg/simulator.py), starts the [DearPyGui UI](positronic/gui/dpg.py), and records into the [Local dataset writer](positronic/dataset/local_dataset.py).

![Data collection GUI](positronic/assets/docs/dc_gui.png)


#### Driving physical robots

Choose the configuration that matches your setup — all presets are defined in [Data collection script](positronic/data_collection.py) using `configuronic`:

```bash
python -m positronic.data_collection real  --output_dir=~/datasets/franka_kitchen
python -m positronic.data_collection so101 --output_dir=~/datasets/so101_runs
python -m positronic.data_collection droid --output_dir=~/datasets/droid_runs
```

Override components inline (e.g. `--webxr=.iphone`, `--sound=None`, `--operator_position=.FRONT`) or add new configs under [Hardware configs](positronic/cfg/hardware/).

#### Teleoperate with a VR headset (Meta Quest / Oculus)

Both sim and real robots can be controlled by position tracking devices, such as VR Helmet (we use Oculus Quest 3, though any VR device will work) or even phone with WebXR-enabled browser.

Here's how to collect the data using VR Helmet:
1. Launch a collector command with `--webxr=.oculus`.
2. Find the host machine’s IP address (`ipconfig getifaddr en0` on macOS, `hostname -I` on Linux) and connect the headset to the same network.
3. In the headset open **Oculus Browser**, navigate to `https://<host-ip>:5005/`, and accept the security warning once (Advanced → Continue). The certificate is generated by [WebXR driver](positronic/drivers/webxr.py).
4. Click **Enter AR** in the page header and approve the permission dialog so the browser can read tracking data.
5. If you started the collector with `--stream_video_to_webxr=<camera_name>`, a floating panel shows the chosen camera feed. `<camera_name>` must match a key in the `cameras` mapping passed to the collector.
6. Controller mappings (handled by [Button mapping helper](positronic/utils/buttons.py)):
   - **Right B** — start/stop recording (audio cues play if sound is enabled).
   - **Right A** — toggle positional tracking to align the controller pose with the robot.
   - **Right stick press** — abort the current episode and send `roboarm.command.Reset()`.
   - **Right trigger** — analogue gripper control (captured as the `target_grip` signal).

#### Use an iPhone as a controller

1. Launch data collection with `--webxr=.iphone` (frontend defined in [WebXR config](positronic/cfg/webxr.py)).
2. On the phone open **XR Browser** (or any WebXR-capable browser) and visit `http://<host-ip>:5005/`.
3. Tap **Enter AR**, grant camera/gyroscope access, and hold the phone upright; the reticle represents the virtual controller.
4. The on-screen HUD from [WebXR phone UI](positronic/assets/webxr_iphone/index.html) provides:
   - **Record** for episode start/stop.
   - **Track**  to toggle positional tracking.
   - **Reset**  to abort the current run and resetting the robot.
   - **Gripper slider** to control the width of your gripper.

#### Collection tips

- Watch the DearPyGui window to confirm cameras and robot state update reliably. Restart an episode instead of saving partial data when something goes wrong.
- Record a few calibration runs first, review them, then capture full demonstrations.

### 3. Review and curate datasets

Browse datasets with the [Positronic Server app](positronic/server/positronic_server.py).

Inspect recordings before training:

```bash
python -m positronic.server.positronic_server \
    --root=~/datasets/stack_cubes_raw \
    --port=5001
```

The viewer loads `LocalDataset` in the background, generates `.rrd` files with the copies of each episode under `~/.cache/positronic/server/`, and serves a Rerun viewer per episode.

![Dataset viewer](positronic/assets/docs/server_screenshot.png)

The UI is read-only for now: mark low-quality runs while watching, then rename or remove the corresponding episode directories manually. Static metadata lives in each episode’s `episode.json` if you need to edit it outside the GUI.

We will support curation mode in the future releases.

### 4. Convert to LeRobot

Use the [LeRobot conversion helper](positronic/training/to_lerobot.py) until native training lands.

Until training scripts consume Positronic datasets directly, convert curated runs into LeRobot format:

```bash
uv run --with-editable . -s positronic/training/to_lerobot.py convert \
    --dataset.path=~/datasets/stack_cubes_raw \
    --output_dir=~/datasets/lerobot/stack_cubes \
    --task="pick up the green cube and place it on the red cube" \
    --fps=30
```

The converter reads your data through `positronic.cfg.dataset.transformed` (see [Dataset config module](positronic/cfg/dataset.py)), applies the same observation/action transforms used at inference time, and writes a `LeRobotDataset`. Re-run the command whenever you tweak transforms or add new episodes. To extend an existing LeRobot dataset:

```bash
uv run --with-editable . -s positronic/training/to_lerobot.py append \
    --dataset_dir=~/datasets/lerobot/stack_cubes \
    --dataset.path=~/datasets/stack_cubes_new
```

Keep the original Positronic datasets — once native training lands you will no longer need this conversion step.

### 5. Train a policy

Run the [LeRobot training driver](positronic/training/lerobot_train.py) with Positronic configs.

Train an ACT policy using LeRobot’s pipeline configured for Positronic observations and actions:

```bash
uv run --with-editable . -s positronic/training/lerobot_train.py \
    --dataset_root=~/datasets/lerobot/stack_cubes \
    --base_config=positronic/training/train_config.json
```

Checkpoints and logs are written under `outputs/train/<timestamp>_<job_name>/`. Adjust the [training script](positronic/training/lerobot_train.py) to change architectures, backbones, or devices. When Positronic-first training is ready you will point the trainer at the raw dataset instead.

### 6. Validate policies

Replay checkpoints through the [Inference script](positronic/run_inference.py) to test in MuJoCo or log hardware runs.

Run the trained policy in MuJoCo, record diagnostics, and optionally stream a GUI:

```bash
python -m positronic.run_inference sim_act \
    --policy.checkpoint_path=~/datasets/lerobot/stack_cubes/checkpoints/last/pretrained_model \
    --device=mps \
    --simulation_time=60 \
    --output_dir=~/datasets/inference_logs/stack_cubes_act \
    --show_gui
```

The [Inference script](positronic/run_inference.py) wires the MuJoCo scene, [Observation encoders](positronic/cfg/policy/observation.py), and [Action decoder](positronic/cfg/policy/action.py). Passing `--output_dir` enables another `DsWriterAgent` so the run can be replayed in the dataset viewer.

---

## Development workflow

Run tests and linters from the root directory:

```bash
uv run pytest --no-cov
uv run flake8
```

Use `uv add` / `uv remove` to modify dependencies and `uv lock` to refresh the lockfile. Contributions should include or update tests when code changes affect behaviour.

## How Positronic differs from LeRobot

[LeRobot](https://github.com/huggingface/lerobot) packages imitation- and reinforcement-learning algorithms, public datasets, and affordable reference robots so anyone can train and share policies. Positronic builds on that progress but is **aimed at teams running professional systems end to end**:

- **Runtime & middleware.** [`pimm`](pimm/README.md) wires sensors, arms, teleop clients, and inference loops on production hardware without ROS launch files or bespoke DSLs.
- **Data operations.** [Positronic dataset library](positronic/dataset/README.md) is a powerful library for robotic datasets.
- **Deployment & supervision.** WebXR teleoperation ([WebXR config](positronic/cfg/webxr.py)), simulator loaders ([MuJoCo transform helpers](positronic/simulator/mujoco/transforms.py)), an expanding set of hardware drivers ([Drivers package](positronic/drivers)), and inference loggers ([Inference script](positronic/run_inference.py)) help you roll out policies, watch live telemetry, and capture diagnostics for regression analysis.

LeRobot is the learning engine; Positronic is the operations stack that keeps data flowing, policies running, and teams iterating on real robots.

## Roadmap

Our plans evolve with your feedback. Highlights for the next milestones:

- **Short term**
  - **Policy presets for $\pi_0$ and GR00T.** Bundle ready-to-use configs in [policy config module](positronic/cfg/policy/policy.py) so switching policies is a single flag.
  - **Automated evaluation harness.** Extend [Inference script](positronic/run_inference.py) and the MuJoCo loaders in [MuJoCo transform helpers](positronic/simulator/mujoco/transforms.py) to score new checkpoints automatically on curated scenarios.
  - **Richer Positronic Server.** Teach [Positronic Server](positronic/server/positronic_server.py) to open transformed datasets, surface static/meta fields, and offer annotation + filtering flows for rapid triage.
  - **PyTorch bridging layer.** Provide a native adapter on top of [Positronic dataset library](positronic/dataset/README.md) so training scripts can stream tensors without an intermediate export.
  - **Direct LeRobot integration.** Let [LeRobot training driver](positronic/training/lerobot_train.py) read Positronic datasets directly, retiring the temporary [LeRobot conversion helper](positronic/training/to_lerobot.py) conversion.
  - **Remote inference primitives.** Add network-friendly transports to [Inference script](positronic/run_inference.py) so policies can live on a different machine than the robot.
- **Medium term**
  - **SO101 leader support.** Promote SO101 from follower mode to a first-class leader arm in [Hardware configs](positronic/cfg/hardware).
  - **New operator inputs.** Ship keyboard and gamepad controllers inside [Drivers package](positronic/drivers) for quick teleop on commodity hardware.
  - **Streaming datasets.** Introduce a cloud-ready dataset backend building on [Local dataset writer](positronic/dataset/local_dataset.py) for long-running data collection jobs.
  - **Community hardware.** Continue adding camera, gripper, and arm drivers requested by adopters.
- **Long term**
  - **Distributed scheduling.** Layer cross-machine orchestration on [`pimm`](pimm/README.md) so one supervisor can coordinate multiple collectors, trainers, and inference nodes.
  - **Hybrid cloud workflows.** Support ingestion of episodes straight into object storage while keeping local curation/debugging fast, plus optional cloud-hosted inference services.
  - **Remote inference in the cloud.** Deploy heavy but capable ML models in the cloud to control and operate your local hardware.

Let us know what you need at hi@positronic.ro or [open a feature request at Github](https://github.com/Positronic-Robotics/positronic/issues/new).
