See the roadmap Join the movement

Open standards for AI agents

Building safe, open agentic AI infrastructure for everyone.

We are a global coalition advancing interoperable tooling, trustworthy governance, and a thriving ecosystem for AI agents.

Join the movement See the roadmap
Jan 2026
Foundation established
3
Structural risks to solve
China → Global
Our positioning
Live Updated weekly

Agentic AI ecosystem in motion

  • Q2

    Open Agent Safety Benchmark

    Community-driven evaluations for autonomy, alignment, and resilience.

  • Q3

    Interoperability Standards

    Reference specs for agent messaging, memory, and tool safety.

  • Q4

    Global Builder Summits

    Multi-city gatherings to publish learnings, roadmaps, and open datasets.

Created with researchers, startups, enterprises, and public institutions.

Our mission

Enable open, safe, and useful agentic AI for society.

We coordinate standards, funding, and governance so that builders across the world can create trustworthy agentic systems—independent, interoperable, and aligned with human values.

Open protocols

We maintain open specifications for agent messaging, identity, safety, and data stewardship so ecosystems can interoperate.

Safety & assurance

We convene researchers to define benchmarks, threat models, and red-team exercises for autonomous systems.

Inclusive governance

We bring civil society, academia, and industry together to shape transparent, accountable governance.

Strategic pillars

What we build together.

From technical standards to public-private collaboration, we focus on the capabilities that make agentic AI reliable, transparent, and universally accessible.

Infrastructure

Reference architecture

Open-source blueprints for memory, planning, and evaluation layers that teams can adopt and extend.

  • Composable agent lifecycle models
  • Secure tool & data access controls
  • Shared telemetry for safety signals
Standards

Interoperability & trust

Specifications and conformance suites that let agent networks collaborate across vendors and jurisdictions.

  • Open agent interface spec (OAIS)
  • Identity, provenance, and auditability
  • Safety, alignment, and resilience benchmarks
Impact

Ecosystem programs

Fellowships, grants, and summits that accelerate open-source agentic AI and real-world deployments.

  • Builder residencies & red-team labs
  • Global policy and safety forums
  • Open datasets & evaluation challenges

Programs

Where we invest.

Our programs channel funding, knowledge, and community energy toward the most urgent needs for agentic AI.

2026

Open Agent Safety Benchmark

Collaborative evaluation suite and public leaderboards for autonomy, reliability, and security.

2026

Agent Interoperability Lab

Reference implementations for messaging, memory exchange, and safety guardrails.

2026

Global Builder Summits

Multi-region gatherings to publish roadmaps and deepen collaboration between industry and researchers.

2026

Open Impact Fund

Non-dilutive grants for open-source agent frameworks, safety tooling, and public-good deployments.

Members

Our member organizations

The Open Agentic AI Foundation would not be possible without the support, development efforts, and open collaboration of our founding platinum members.

Guidance Agency

Our guidance organizations

Strategic guidance and oversight from leading organizations ensuring our mission aligns with global standards and best practices.

Partner Communities

Our partner communities

Collaborative communities working together to advance open agentic AI standards and innovation.

ANP社区
CO-SIGHT社区
CO-TAP社区
OpenTenBase社区
RAGFlow社区
WeOpen社区

Community

A global coalition of builders, researchers, and policymakers.

We convene open-source contributors, startups, enterprises, universities, and civil society to co-design the next era of agentic AI.

Open & neutral

We publish artifacts under open licenses and encourage cross-organizational governance.

Safety first

Threat modeling, evals, and incident response are built into every program.

Global access

We support multilingual resources and regionally distributed summits.

Resources

Start building with us.

Explore our living roadmap, specifications, and community channels to contribute from day one.

Roadmap

Agent standards preview

Draft specifications for interfaces, memory, and safe tool access.

View the docs
Community

Join the forum

Connect with contributors, propose initiatives, and join working groups.

Open workspace
Funding

Apply for grants

Support for open-source projects, research, and deployments that move the ecosystem forward.

Start an application

Get involved

Bring your expertise to The Open Agentic AI Foundation.

Contribute to standards, participate in safety evaluations, and shape open governance for agentic systems.

Join the coalition Browse resources