About Project NANDA¶

Networked AI Agents in Decentralized Architecture (NANDA) is an open-source initiative and global consortium building the foundational infrastructure for the Internet of AI Agents — an open, decentralized ecosystem where trillions of autonomous AI agents can discover each other, collaborate, communicate, and transact across organizational boundaries.

Origin & Academic Heritage¶

Project NANDA originated from research at the MIT Media Lab (Camera Culture research group led by Prof. Ramesh Raskar). It builds upon foundational academic breakthroughs in AutoML, split learning, privacy-preserving machine learning, and multi-agent coordination.

Today, Project NANDA operates as a global collaborative effort uniting researchers, engineers, academic institutions, and industry leaders from MIT, Stanford, Google, Cisco, OpenAI, Dell Technologies, Intel, AWS, and civic organizations worldwide.

The Core Challenge: Why the Agentic Web Needs New Infrastructure¶

The existing Internet infrastructure was engineered in the 1980s to serve static HTML documents and human-facing web browsers. As artificial intelligence advances from passive copilots to autonomous, proactive agents executing complex goals on behalf of users and organizations, the legacy web faces four fundamental choke points:

Discovery (Beyond DNS):
The Domain Name System (DNS) was designed to map static hostnames to IP addresses with minute-to-hour propagation delays. Autonomous agents require sub-second global discovery, capability querying, and real-time endpoint resolution across dynamic edge and cloud environments.
Identity & Attestation (Beyond Certificate Authorities):
Traditional Certificate Authorities (CAs) only verify that an entity controls a domain name. They cannot verify an agent’s behavioral constraints, operational authority, cryptographic capability attestations, or execution history.
Orchestration & Interoperability:
The emerging multi-agent landscape is rapidly fragmenting into walled-garden platforms and incompatible frameworks (MCP, A2A, HTTPS, LangChain, CrewAI). Without a neutral resolution and translation layer, agents cannot collaborate across organizational and technological boundaries.
Economic Settlement & Value Exchange:
The web lacks native primitives for autonomous software entities to dynamically quote, price, meter, and settle micropayments for data, computation, and specialized skills.

Project NANDA addresses these challenges by building open, neutral, and decentralized infrastructure layers that make the agentic web scalable, secure, and permissionless.

The Three Pillars: Our Approach¶

Project NANDA employs a three-pronged approach to catalyze the Internet of AI Agents:

1. Technology & Open Standards¶

We develop open-source protocols, reference implementations, and developer SDKs through a global collective of researchers and engineers: - NANDA Index: A lightweight, decentralized discovery directory operating as the universal "phone book" for AI agents. - AgentFacts / host39: Verifiable, cryptographically signed JSON-LD passports that declare an agent's capabilities, credentials, and endpoints. - NANDA Adapter & SDK: Multi-protocol translation bridges that connect agents across disparate frameworks including Model Context Protocol (MCP), Agent-to-Agent (A2A), and HTTPS. - NEST TestNet & Nanda Town: Shared simulation sandboxes and protocol test tracks for multi-agent experimentation.

We believe the future of computing must remain open, equitable, and safe: - Anti-Monopoly & Neutrality: Ensuring the agentic web is not captured by centralized gatekeepers or proprietary platforms. - Privacy by Design: Implementing zero-knowledge, dual-path resolution mechanisms that allow agents to discover services without leaking sensitive query patterns or business intent. - Safety & Ethics: Establishing verifiable guardrails, parental controls, and auditing frameworks to protect users, children, and vulnerable communities.

3. Venture Ecosystem & Industry Coalition¶

We bridge lab breakthroughs into production-grade market adoption: - Fostering an international coalition of enterprise partners, startups, and open-source contributors. - Supporting early-stage venture incubation and developer fellowships to accelerate real-world deployments. - Stimulating public sector and civic adoption to empower open digital public infrastructure (DPI).

Architectural Evolution: Three Developmental Stages¶

Project NANDA's technical roadmap progresses through three interconnected phases:

Stage 1: Foundations of the Agentic Web
Establishing the foundational discovery and communication rails. This encompasses the NANDA Index, AgentFacts schemas, cross-platform protocol adapters (A2A, MCP, HTTPS), and developer onboarding SDKs.
Stage 2: Agentic Commerce
Deploying economic primitives and decentralized resource markets. This includes dynamic knowledge pricing algorithms, edge AI integration for local hardware gateways, autonomous payment and escrow protocols, and compute/data trading networks.
Stage 3: Society of Agents
Enabling population-scale collective intelligence. This includes Large Population Models (LPMs), federated collaborative learning across institutional data silos without compromising privacy, and cross-organizational multi-agent swarms.

Institutional Home & Governance¶

Project NANDA is governed under the Foundation for Agentic Networks (FAN), a neutral, public-benefit non-profit organization dedicated to advancing open-agent-network standards.

FAN provides neutral stewardship, transparent intellectual property management, and sustained governance to guarantee that core agent infrastructure remains an open public good accessible to all innovators.

Open Source & Permissive Licensing¶

All Project NANDA specifications, reference implementations, and research software are released under permissive open-source licenses (Apache 2.0 and MIT).

We welcome open-source contributions from developers, researchers, and enterprises worldwide via our GitHub repositories.

Global Community & Footprint¶

Project NANDA is supported by a vibrant and rapidly expanding international ecosystem:

10,000+ Community Members & Builders
60+ Academic & Corporate Collaborations
15+ Foundational Research Papers
150+ Technical Webinars & Workshops
15+ Global Hackathons & Residencies
20+ Regional Chapter Locations worldwide

Get Involved¶

Whether you are a developer building autonomous agents, a researcher investigating multi-agent systems, an enterprise deploying agentic workflows, or an investor supporting the ecosystem, there are many ways to participate:

Explore Research: Read our latest papers in the Research Publications catalog.
Contribute to Research: Join our interdisciplinary working groups via the Writing Group Guide.
Build with NANDA: Explore our open-source tools on GitHub and start building with the NANDA Adapter.
Attend Events: Join our weekly community calls, seminars, and technical workshops on Lu.Ma.
Start a Local Chapter: Learn how to lead an agentic community in your city on the Start a Chapter page.
Institutional Partnerships: Contact our core team at contact@projectnanda.org to discuss testbeds, enterprise pilots, or fellowship sponsorships.