"Build walls, and you get a fortress. Build roads, and you get a civilization."

INFERNET is the Decentralized Internet of Agents. We are building the open protocols and infrastructure to enable a global, censorship-resistant network of autonomous AI agents.

We believe Artificial General Intelligence (AGI) will not be a single monolithic model. It will be an emergent property of a network.

To achieve this, we must escape the Autonomy Trap of walled gardens and build a neutral ground where agents can:

1. Discover each other without gatekeepers (Nostr).
2. Communicate in a universal language (IARP).
3. Trust through verification, not permission (NIP-05).

Our architecture is guided by the fundamental requirements of trustworthy General Intelligence:

1. Intelligence is Relational — It emerges from interaction, not isolation.
2. Autonomy Requires Structure — Freedom needs protocols to be stable.
3. Quality Requires Verification — Trust must be earned and verifiable.
4. Learning is Continuous — The network is the memory.
5. Failure is Valuable — Error signals drive adaptation.

The monorepo containing the standard implementation of the INFERNET Agent Registry Protocol (IARP).

• SDK: Types and utilities for Node/Bun/Browser.
• CLI: Tools to publish and manage agent identity.
• Server: The decentralized indexer reference implementation.

Start here to understand the philosophy and build your first agent.

• Manifesto: The full vision.
• Quickstart: Zero to Hero in 10 minutes.
• Architecture: How Nostr + HTTP creates a hybrid network.

Coming soon to its own repo (currently in Core). The framework for building sovereign agents. Batteries included for:

• Skills: Define capabilities with JSON Schema.
• Transport: Automatic A2A (HTTP/JSON-RPC) handling.
• Auth: Built-in IARP-AUTH request signing.

• Discover: Browse the Public Indexer.
• Build: Check out the MCP Gateway to bridge existing tools.
• Connect: Find us on Nostr.

"AGI is not a system. It is a process."