Standards

Standards for the open agentic web.

The foundation designs and contributes to open standards — for the agentic web, for holding AI systems accountable and transparent, and for the governance, safety, and interoperability of the technologies people depend on.

01
The open agentic web
Protocols and specifications for how autonomous systems discover one another, exchange capabilities, transact, and coordinate — without depending on closed platforms or proprietary gateways.
02
Oversight of autonomous agents
As agents take on critical decisions, those decisions need to remain under meaningful oversight. Standards work here covers intervention points, escalation paths, and the controls that keep agentic systems answerable.
03
Accountability & transparency
Disclosures, model and system cards, evaluation reporting, and incident frameworks that make AI systems legible to the people they affect and to the bodies that oversee them.
04
Governance frameworks for AI
Open frameworks for how AI systems are evaluated, released, and held accountable — covering release reviews, third-party audits, and the structures that make independent oversight tractable.
05
Safety, interoperability, accessibility
Cross-cutting standards work — auditable safety methods, interoperable evaluation formats, and accessibility requirements that travel across models and modalities.

Active specifications

OAW · Open Agentic Web StandardsIncubating

A set of open specifications for the agentic web — the layer of the Internet where autonomous AI agents discover, communicate with, transact with, and verify each other.

Where the World Wide Web standardized hypertext (HTML), naming (DNS), and transport (HTTP) so anyone could publish a website, OAW standardizes the equivalents for agents.

github.com/tenzro-foundation/oawRepository →

More forthcoming

Additional working group charters, specifications, and external participation will be listed here as the founding membership is formed and the public agenda is set.