NANDA logo

NANDA

08 agent-communication/NANDA (MIT Media Lab)/ACTIVE/aka Project NANDA, Internet of AI Agents

NANDA is MIT Media Lab's 'Internet of AI Agents' research program — a registry and index layer plus protocols for discovering, verifying, and composing agents at internet scale.

SEC.01

What it does

  • 01Agent registry and index
  • 02Capability search
  • 03Verifiable agent facts
  • 04Research testbeds (AgentFacts, MCP index)
SEC.02

Who it's for

Researchers and platform teams thinking about internet-scale agent discovery beyond any single vendor's directory.

PRICING

Pricing

Free (research project). List prices as checked against the vendor's pricing page on 2026-09-25; plans change often, so confirm before deciding.

caveat · before you adopt

A research program, not a product — pieces are prototypes; the adoption path runs through standards work rather than SDKs.

ALT

Other agent communication platforms (12)

FAQ

Frequently asked

What is NANDA?

NANDA is MIT Media Lab's 'Internet of AI Agents' research program — a registry and index layer plus protocols for discovering, verifying, and composing agents at internet scale.

Who is NANDA for?

Researchers and platform teams thinking about internet-scale agent discovery beyond any single vendor's directory.

How much does NANDA cost?

Free (research project). List prices as checked against the vendor's pricing page on 2026-09-25; plans change often, so confirm before deciding.

Does NANDA run locally or in the cloud?

NANDA runs locally on your machine and in the vendor's cloud. Its interfaces are an API or protocol.

What are alternatives to NANDA?

Other agent communication platform entries in the catalog: OpenScout / Scout, A2A Protocol, AGNTCY, Agent Network Protocol, BeeAI, CrewAI, Microsoft Agent Framework, LangGraph, MetaGPT, CAMEL, Agency Swarm, OpenAI Agents SDK. Different categories solve the same problem differently — see the Agent communication & collaboration category page for tradeoffs.

What should I know before adopting NANDA?

A research program, not a product — pieces are prototypes; the adoption path runs through standards work rather than SDKs.