Engineering a New Trust Layer for the Internet

The internet connects people, organizations, communities, devices, and AI agents — but turning those connections into collaboration takes trust. Yet the internet was built without a trust layer — and the rise of agentic AI has made that gap urgent. That’s why we build the P2P Trust Stack: v0.2.0 is released and open source — tools that let participants communicate directly and establish trust without a central database. We also help teams put the stack to work — from architecture and standards guidance to prototypes and integration.

Explore the P2P Trust Stack Discuss your project Developer documentation

v0.2.0 — The Open Source P2P Trust Stack Is Here

The core protocol engine and the application-facing agent API are now open source and published to npm. v0.2.0 is a working implementation of decentralized identifiers, peer-to-peer relationships, and verifiable credentials that anyone can build on — under GPL-3.0, or a commercial license.

📦

Two Packages, One Stack

@p2p-trust-stack/core is the protocol engine — cryptography, DIDs, the encrypted keystore, TSP, libp2p transport, relationships, VTCs/VRCs, and the transaction coordination layer. @p2p-trust-stack/app-api is the application-facing agent API (createAgentStack(), the typed TrustStack domains, and a closed event catalogue) and depends on core.

# application developers
npm install @p2p-trust-stack/app-api

# protocol-only embedders
npm install @p2p-trust-stack/core

core on npm ↗ · app-api on npm ↗

🔓

Open Source, Dual-Licensed

The full stack is public — source, tests, schema migrations, CI, and governance docs. Use it under GPL-3.0-only, or under a commercial license when GPL-3.0 obligations are not acceptable. All contributors sign the Contributor License Agreement.

GitHub repository ↗ · CHANGELOG ↗ · Commercial license ↗

📚

Documented and Demonstrated

The API Reference documents every method of the published packages — 8 groups, 117 entries, grouped by the four Trust over IP layers with a module-by-module core reference. The stack already drives real flows: VTA creation and registration, agent names, confidential P2P chat, mutual acknowledgment, VTC creation and governance, and credential issuance through the Credential Factory.

API Reference → · Developer Docs →

What Trust Stackers Does

We engineer trust infrastructure in two complementary ways:

🧱

We Build the P2P Trust Stack

The stack gives people and agents their own identifiers, records their relationships, and lets them communicate in private — without a central database. v0.2.0 is released and open source: install it from npm and build on it.

The Stack →

🧭

Engineering and consulting

We build the P2P Trust Stack, and we help others put it to work. Bring us your use case: we’ll help assess the fit, prototype an agreed workflow, and integrate or extend the relevant stack components.

Discuss your project →

How the pieces fit together

An application may need to know who issued a claim, whether a participant is a community member, or whether a software agent may act on someone’s behalf.

Decentralized identifiers provide references to participants. Cryptographic mechanisms support authentication and signature verification. Verifiable credentials carry claims that an application can verify and evaluate under its own rules.

Taken together, these form a decentralized trust graph: each participant holds their own node, and relationships are the signed edges between them — no single database stores the whole picture.

For example, a community can issue a membership credential that a service verifies when deciding whether to grant access.

Verifying a credential’s proof does not, by itself, establish that every claim is true or that its issuer should be trusted. This is an illustrative example — what an application does with a verified credential is policy, not cryptography.

Watch the existing demo → · Explore the architecture →

The P2P Trust Stack

The stack is organized around the four-layer Trust over IP (ToIP) model. Here's where each layer stands today.

Layer	Name	Technology	Status
4	Trust Applications PNM / CNM / Developer Apps	End-user UI/UX for creating, joining, and managing trust applications, communities, and networks. The Agent App reference applications (web + TUI, plus a mobile Android/iOS example) are built and operational — registering VTAs, founding and administering VTCs, issuing VRCs, managing community rules, sending invitations, and chatting — and the stack is already being embedded in other test applications.	Active — Reference Apps Built
3	Trust Tasks VTCs & VRCs	W3C Verifiable Credentials issued through the unified Credential Factory · data-driven Governance Guard and Policy Engine · trust-task ceremonies · VTC establishment, membership and mutual acknowledgment flows operating · released in v0.2.0 · trusttasks.org (interop planned)	Foundations Laid
2	Trust Spanning VTAs & TSP	Trust Spanning Protocol (TSP) · DID-based identifiers · libp2p transport · authenticated messaging · optional end-to-end encryption · peer relationship establishment · agent name registration — released in v0.2.0	Active — Released in v0.2.0
1	Trust Support DID Infrastructure	did:scid:vh (self-certifying, hash-chained, operational) · native key rotation with witness attestation — released in v0.2.0	Active — Released in v0.2.0

Full Stack Details → What's Next — Roadmap →