AI Origins · Network Architecture

Core, fixed, mobile, reserve — a protocol-led node network built around identity, transport, and trust.

AI Origins is designed as a specification network rather than a hardware sales network. The core idea is simple: publish the rules, define the node roles, standardize the identity and transport interface, and let compliant operators join through protocol compatibility and trust-aware admission. Existing audit files and folders can remain untouched if they are not in conflict; the architecture grows by layering, not by destructive rearrangement.

Core principle

Spec-first

Implementation may vary. Identity, endpoint correctness, protocol compatibility, and admission policy remain fixed.

Transport model

10.66.66.x

Private mesh transport based on WireGuard, with role-based address ranges and future room for growth.

Trust model

Audit layer

Join, service, and trust stay separated: onboarding is one layer, audit is another, core authority remains tighter.

Recommended site structure

Public structure

https://aiorigins.org/
https://aiorigins.org/join/
https://aiorigins.org/spec/
https://aiorigins.org/rules/
https://aiorigins.org/verify/
https://aiorigins.org/nodes/
https://aiorigins.org/audit/
https://aiorigins.org/node.json
https://aiorigins.org/portal.json
https://aiorigins.org/agent.json

Layer meaning

  • /join/ — public onboarding entry
  • /spec/ — architecture and protocol definition
  • /rules/ — admission and role policy
  • /verify/ — Join Check explanation or interface
  • /nodes/ — node categories and examples
  • /audit/ — trust and evidence layer
Your existing audit-era folders and files do not need to be moved merely for neatness. If they do not conflict with the new layout, leave them in place. The safer rule is: preserve working trust-layer files, add new onboarding/spec pages alongside them, and avoid touching anything that may break a stable path.

Node role model

CORE

Anchor node. Governance, protocol source, identity root, audit root, and admission authority.

core-londonhigh trustprivate control

FIXED

Long-running production nodes with stable location and durable connectivity. Suitable for service execution, regional support, and later quorum evolution.

fixed-qingdaofixed-helsinkifixed-vienna

MOBILE

Portable presence node. Temporary online proof, witness logging, travel access, and light participation without carrying core production duties.

mobile-evanpresencewitness

RESERVE / EDGE / SERVICE

Reserve nodes wait for deployment or replacement. Edge nodes are public-facing access points. Service nodes provide narrower scoped functions under policy constraints.

reserve-euedgeservice

Current Eurasian architecture

Minimum stable shape

                 core-london
                (London VPS)
                     │
          ┌──────────┼──────────┐
          │                     │
   fixed-helsinki         fixed-qingdao
     (Finland VPS)         (planned node)

This is the first meaningful distributed shape: a core plus two fixed nodes. Once the third node is active, the structure stops being a mere tunnel pair and begins to act like a real network.

Expanded corridor

                 core-london
                      │
        ┌─────────────┼─────────────┐
        │             │             │
 fixed-helsinki   fixed-vienna   fixed-qingdao
        │             │             │
 fixed-amsterdam fixed-istanbul fixed-incheon
                                      │
                                 fixed-fukuoka
                                      │
                                  mobile-evan

This corridor is not required all at once. It shows the likely growth direction from Europe to East Asia, with mobile presence orbiting the fixed backbone.

Transport plan

Address convention

10.66.66.1      core-london
10.66.66.10-19  fixed nodes
10.66.66.20-29  mobile nodes
10.66.66.30-39  reserve nodes
10.66.66.40-49  edge/service nodes

WireGuard peer model

[Interface]
PrivateKey = YOUR_PRIVATE_KEY
Address = 10.66.66.10/32
DNS = 1.1.1.1

[Peer]
PublicKey = CORE_LONDON_PUBLIC_KEY
Endpoint = aiorigins.org:51820
AllowedIPs = 10.66.66.0/24
PersistentKeepalive = 25

Identity model

Node identity

{
  "schema": "https://aiorigins.org/spec/node/v1",
  "network": "AI Origins",
  "policy_version": "Join Node Spec v1.0",
  "node_id": "fixed-qingdao-01",
  "node_role": "fixed",
  "node_label": "Qingdao Fixed Node",
  "operator": {
    "name": "Evan Bei",
    "contact": "[email protected]"
  },
  "region": "CN-QD",
  "transport": {
    "type": "wireguard",
    "wg_address": "10.66.66.10/32",
    "endpoint": "qingdao.example.net:51820",
    "public_key": "BASE64_PUBLIC_KEY"
  },
  "capabilities": ["presence", "rpc", "witness"],
  "admission": "reviewed",
  "updated_at": "2026-03-06T00:00:00Z"
}

Why identity matters

  • It keeps hardware optional and protocol mandatory.
  • It lets operators implement by their own method.
  • It gives a consistent machine-readable declaration for trust checks.
  • It separates node naming, endpoint definition, and admission status.

Trust and onboarding layers

Onboarding flow

Read rules
→ choose node role
→ prepare node.json + public key + endpoint
→ self-deploy by any compliant method
→ run Join Check
→ PASS / WARN / FAIL
→ admission review if needed
→ join network

Join Check API sketch

POST /api/join-check
Content-Type: application/json

{
  "network": "AI Origins",
  "policy_version": "Join Node Spec v1.0",
  "node_id": "fixed-qingdao-01",
  "node_role": "fixed",
  "region": "CN-QD",
  "endpoint": "qingdao.example.net:51820",
  "public_key": "BASE64_PUBLIC_KEY",
  "metadata_url": "https://example.org/node.json"
}

Response 200
{
  "ok": true,
  "status": "PASS",
  "checks": {
    "metadata": "PASS",
    "public_key": "PASS",
    "endpoint": "WARN",
    "policy": "PASS"
  },
  "message": "Node is protocol-compatible and may proceed to admission review."
}

Operational rule for existing audit files

Existing files and folders created during earlier audit work can remain where they are if they do not conflict with the new onboarding and specification pages. There is no benefit in moving stable trust-layer files merely for visual tidiness. The safer operational rule is:

Do not touch working audit files unless there is a direct need.
Add new pages alongside the working structure.
Treat audit as the trust layer.
Treat join/spec/rules/verify as the onboarding layer.
Preserve stable paths first; optimize later.