Physical Intelligence

Physical Intelligence (often styled "Pi" or "π") is a San Francisco-based research company building general-purpose foundation models for robotics with the stated goal of producing learning algorithms that can control any robot to do any task. The company has published a continuing line of Vision-Language-Action (VLA) models: π0 (October 2024, first generalist multi-task multi-robot policy), π0-FAST (autoregressive variant via Real-time Action Chunking / FAST tokenization), π0.5 (April 2025, open-world generalization), π*0.6 (November 2025, reinforcement-learning from experience), and π0.7 (April 2026, steerable model with emergent capabilities). Physical Intelligence releases significant work as open source: the openpi repository (~12K stars) is the canonical home for π0 weights and code, with companion repos including real-time-chunking-kinetix, pi-data-sharing, aloha, augmax, and rlds_dataset_builder. The company does not yet offer a hosted commercial API; access to the platform is via open-weight models and research collaborations.

Physical Intelligence is profiled on the APIs.io network. Tagged areas include Robotics, Foundation Models, Vision Language Action, Embodied AI, and Reinforcement Learning.

Physical Intelligence’s developer surface includes engineering blog and 9 more developer resources.

🌐 Visit website 📡 Source on GitHub

7.9/100 minimal ▬ flat Agent 0/100 human only Full breakdown ↓
scored 2026-07-27 · rubric v0.5

0 APIs

RoboticsFoundation ModelsVision Language ActionEmbodied AIReinforcement LearningImitation LearningOpen SourceOpen Weightspi0openpiManipulationGeneralist Policy

Kin Score

Kin Score How this is scored →
scored 2026-07-27 · rubric v0.5

Composite quality — 7.9/100 · minimal

Contract Quality 0.0 / 25

Developer Ergonomics 0.4 / 20

Commercial Clarity 0.0 / 20

Operational Transparency 0.7 / 13

Governance 0.0 / 12

Discoverability 6.8 / 10

Agent readiness — 0/100 · human only

Machine-Readable Contract 0 / 18

Agentic Access Contract 0 / 15

MCP Server 0 / 12

Machine-Readable Auth 0 / 10

Idempotency 0 / 9

Stable Error Semantics 0 / 8

Request/Response Examples 0 / 7

Rate-Limit Signaling 0 / 7

Typed Event Surface 0 / 6

Agent Skills 0 / 5

Well-Known Catalog 0 / 4

Consent & Bot Identity 0 / 3

Improve this rating by publishing the missing artifacts — every area above can be raised, and the full rubric is at apis.io/rating/. This rating is computed from github.com/api-evangelist/physical-intelligence: open an issue to ask a question, or submit a pull request to add artifacts. Want it done for you? Prioritized profiling — $2,500 →

Build 1

SDKs, sample code, and the tooling you integrate with

GitHubOrganization

GitHubOrganization

Access & Security 1

Authentication, authorization, and security posture

DomainSecurity

DomainSecurity

Company 6

The organization behind the API

Website

AlternateWebsite

Blog

Careers

Twitter

LinkedIn

Other 2

Properties that don't map to a standard resource type

Research

Research

OpenPi

OpenPi

Source (apis.yml)

aid: physical-intelligence
name: Physical Intelligence
description: 'Physical Intelligence (often styled "Pi" or "π") is a San Francisco-based research company building general-purpose
  foundation models for robotics with the stated goal of producing learning algorithms that can control any robot to do any
  task. The company has published a continuing line of Vision-Language-Action (VLA) models: π0 (October 2024, first generalist
  multi-task multi-robot policy), π0-FAST (autoregressive variant via Real-time Action Chunking / FAST tokenization), π0.5
  (April 2025, open-world generalization), π*0.6 (November 2025, reinforcement-learning from experience), and π0.7 (April
  2026, steerable model with emergent capabilities). Physical Intelligence releases significant work as open source: the openpi
  repository (~12K stars) is the canonical home for π0 weights and code, with companion repos including real-time-chunking-kinetix,
  pi-data-sharing, aloha, augmax, and rlds_dataset_builder. The company does not yet offer a hosted commercial API; access
  to the platform is via open-weight models and research collaborations.'
type: Index
accessModel:
  pricing: unknown
  onboarding: unknown
  trial: false
  try_now: false
  public: false
  label: Unknown
  confidence: low
  source: []
  generated: '2026-07-22'
  method: derived
position: Provider
access: 3rd-Party
image: https://kinlane-images.s3.amazonaws.com/shared/apis-json/icons/physical-intelligence.png
tags:
- Robotics
- Foundation Models
- Vision Language Action
- Embodied AI
- Reinforcement Learning
- Imitation Learning
- Open Source
- Open Weights
- pi0
- openpi
- Manipulation
- Generalist Policy
url: https://raw.githubusercontent.com/api-evangelist/physical-intelligence/refs/heads/main/apis.yml
created: '2026-05-23'
modified: '2026-05-23'
specificationVersion: '0.20'
apis: []
common:
- type: DomainSecurity
  url: security/physical-intelligence-domain-security.yml
- type: Website
  url: https://www.physicalintelligence.company
- type: AlternateWebsite
  url: https://www.pi.website
- type: Blog
  url: https://www.pi.website/blog
- type: Research
  url: https://www.pi.website/research
- type: GitHubOrganization
  url: https://github.com/Physical-Intelligence
- type: OpenPi
  url: https://github.com/Physical-Intelligence/openpi
- type: Careers
  url: https://www.pi.website/careers
- type: Twitter
  url: https://twitter.com/physical_int
- type: LinkedIn
  url: https://www.linkedin.com/company/physical-intelligence
maintainers:
- FN: Kin Lane
  email: kin@apievangelist.com