Intelligence Operating Network System · v0.3

Intelligence from
traversal, not weights

An open protocol for publishing, connecting, and traversing typed knowledge claims. Any model can query it. Any node can join it. The network is the durable asset.

Query the network → Run your own node
—
Published CBBs
—
Typed Relationships
—
NSI Clusters
—
Active Nodes
The protocol

Three primitives.
Infinite composition.

IONS externalizes intelligence into a traversable network of atomic knowledge claims. Models are replaceable interpreters. The network is the durable asset.

⬡
Cognitive Building Block
A single atomic claim with evidence, confidence, provenance, and a cryptographic hash. Not a document chunk. A typed, addressable knowledge artifact.
⟶
Typed Relationship
Connects CBBs with semantic type — supports, contradicts, causes, depends_on, extends. Relationships are first-class artifacts with confidence and rationale.
◎
Reasoning Path
The stored result of traversal. Not discarded after the query — saved as a reusable artifact with path confidence, evidence chain, and every step visible.
⊕
NSI Cluster
Narrow Super Intelligence groups emerge from semantic clustering of CBB domains. No manual taxonomy — clusters form naturally as contributors add knowledge.
⬡⬡
Federation
Any node implementing the protocol can register and participate. Queries traverse all connected nodes automatically, merging ranked paths across the network.
🔒
Integrity
Every CBB and relationship has a canonical SHA-256 hash. Changes are detectable. Evidence references can be private while claims are public.
Query flow
❓
Query
Natural language
→
🔍
Discover
Starting CBBs
→
⬡
Traverse
Depth-first, max 5 hops
→
📊
Score
CBB × REL × Evidence
→
✨
Synthesize
Grounded answer
→
📋
Path
Stored artifact
Live demo

Ask the genesis network

Queries traverse 6,000+ typed knowledge claims across 24 NSI clusters. Every answer shows its reasoning chain.

How does institutional memory compound competitive advantage?
Why do AI pilots succeed but fail to scale?
What makes knowledge reusable across an organization?
How does proof of work create scarcity in Bitcoin?
3 queries per minute · 10 per hour · Genesis node · run your own node for unlimited access
Differentiation

How IONS differs from
RAG and knowledge graphs

RAG
✗ Chunks discarded after query
✗ No typed relationships
✗ No path confidence score
✗ No evidence lineage
✗ No reusable reasoning paths
~ Retrieves relevant context
Knowledge Graph
✗ No path artifact lifecycle
✗ No confidence scoring
✗ No evidence references
✗ No creator attribution
~ Typed edges
~ Graph traversal
IONS
✓ Reasoning paths stored as artifacts
✓ 8 typed relationship types
✓ CBB × REL × Evidence confidence
✓ Cryptographic integrity
✓ Federated multi-node traversal
✓ Open protocol, MIT license
Get started

Three ways to participate

01 — QUERY
Use the genesis node
Query 6,000+ CBBs across 24 NSI clusters via the demo above or the full explorer. No account required.
Open explorer →
02 — RUN
Run your own node
Clone the repo, add your OpenRouter key, and start your node in minutes. Contribute your domain knowledge to the network.
git clone github.com/nomad505050/ions-genesis cp .env.example .env docker compose up -d
03 — FEDERATE
Join the network
Register your running node with the genesis node. Your CBBs immediately participate in federated traversal across all connected nodes.
curl -X POST \ http://162.243.203.243:8000/nodes/register \ -d '{"node_id":"your_id", "public_api_base":"https://..."}'