# Robbyant/lingbot-va

[RSS 2026] Causal video-action world model for generalist robot control

Repository: https://github.com/Robbyant/lingbot-va
Canonical: https://ross.abutalabs.com/products/lingbot-va
Homepage: https://technology.robbyant.com/lingbot-va
Language: Python
License: Apache-2.0
License Family: permissive
Last push: 2026-07-09T16:02:38+00:00

## Health v2 (maintenance only)
Score: 56/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 91, release rhythm 35, longevity 15
- inputs: {"age_days": 216, "days_push": 55, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1806, forks 164 (observed 2026-08-28T04:05:39.092101+00:00)

## What it is
LingBot-VA is an autoregressive diffusion framework that unifies video world modeling and robot policy learning in a single interleaved video-action sequence. It provides pretrained weights, inference, and post-training code for generalist robot manipulation control.

## Use cases
- train a video-action world model for robot manipulation
- evaluate a generalist robot policy on LIBERO and RoboTwin 2.0 benchmarks
- post-train a robot foundation model on custom demonstration data
- predict future video frames and actions for closed-loop robot control
- run image-to-video-action generation for long-horizon manipulation tasks

## When to choose
- you need a state-of-the-art generalist manipulation policy with world-model reasoning
- you want to post-train a pretrained video-action model on your own robot datasets
- you need long-horizon, precision, or deformable-object manipulation with strong generalization

## When to avoid
- you need a lightweight policy for low-compute embedded robot hardware
- you only need classical motion planning or control without learned models
- you lack GPU resources for large diffusion model inference and training

## Facets
- artifact type: library
- maturity: active
- function: machine-learning, deep-learning, llm-inference, simulation, robotics
- domain: robotics, machine-learning, deep-learning, artificial-intelligence, simulation
- platform: python
- tags: world-model, video-action-model, robot-control, diffusion-policy, autoregressive, embodied-ai, manipulation, libero, robotwin, linux, gpu, docker

## Member repositories
- Robbyant/lingbot-va (main) score 56

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:05:39.092101+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T03:21:28.106851+00:00, confidence not recorded.
  - readme: https://github.com/Robbyant/lingbot-va (fetched 2026-08-28T04:05:39.092101+00:00, sha 34150064f383)
  - homepage: https://technology.robbyant.com/lingbot-va (fetched 2026-08-29T11:00:37.274434+00:00, sha f64cd983249c)
  - site_page: https://www.robbyant.com/about-robby (fetched 2026-08-29T11:00:37.278284+00:00, sha 4a2374d8a183)
- Data as of 2026-08-30T08:39:29.467469+00:00.