Encode natural language text into a GENOM protocol message with intent detection and semantic tokenization.

Decode a list of GENOM token IDs back to natural language text. This is the primary and fastest decode method.

Decode a space-separated hex sequence to text. Slower than ID-based decode as it requires dictionary lookup by hex code.

Parse and decode a complete GENOM message (the output format from /encode). Extracts header, intent, payload and validates structure.

Get GENOM dictionary statistics including total token count, sector distribution, and version info.

Search the GENOM dictionary by canonical word name. Supports filtering by sector.

Get detailed information for a specific token by numeric ID or canonical word. Returns complex coordinates, sector, hex code.

Returns metadata about the immutable Genesis 2000 token core vocabulary as defined in GENOM 1.0.

{
  "layer": "genesis",
  "version": "1.0.0",
  "token_count": 2000,
  "frozen": true,
  "description": "The immutable 2000-token core vocabulary as published in GENOM 1.0",
  "domains": {
    "Q1": "Technical (0-90 deg)",
    "Q2": "Entity (90-180 deg)",
    "Q3": "Action (180-270 deg)",
    "Q4": "Abstract (270-360 deg)"
  },
  "source": "https://genom.genisys.online/release/v1/genesis.json",
  "book": "https://www.amazon.com/dp/B0GCN4SF93"
}

Returns all 2000 Genesis tokens with their coordinates on the complex semantic plane. Each token includes token_id, canonical form, sector, hex code, and (r, real, imag) coordinates.

{
  "layer": "genesis",
  "count": 2000,
  "tokens": [
    {
      "token_id": "1",
      "canonical": "the",
      "sector": "abstract",
      "hex_code": "0x0001",
      "r": 1.0,
      "real": 0.0,
      "imag": 1.0
    },
    ...
  ]
}

Register a node into the SKYNET swarm. Returns a JWT session token valid for 30 days. This is the primary way to authenticate.

Send a heartbeat signal to indicate the node is alive. Call every 60 seconds. Requires authentication.

List all active nodes in the SKYNET swarm. Requires authentication.

Download the dna.genom identity file for a node. Returns the parsed YAML content and raw text from S3.

Register a new GENOM instance using a voucher code (from the book or alpha program). Returns access credentials and swarm position.

Check registration status by SHA256 hash of the email address (or first 16 characters).

Report unmapped words encountered during encoding. Called automatically by the encoder, but can also be called manually.

List word candidates pending promotion, sorted by frequency.

Promote a candidate word into the live GENOM dictionary. Computes complex coordinates and assigns to a quadrant. Requires NOD1, NOD2, or GENOM authentication.

Get learning pipeline statistics: total candidates, status distribution, and auto-promote threshold.

Send a GENOM-encoded message from one registered node to another. Messages are stored with 7-day TTL and support intent vectors.

// Request
{
  "sender_id": "9f86d081884c7d65",
  "target_id": "b9e086ffd9a9af44",
  "token_ids": [1, 42, 100, 2000, 500000],
  "intent": "μ",
  "session_id": "conversation-001"
}

// Response
{
  "message_id": "52190bd5-9bab-4a69-...",
  "status": "pending",
  "timestamp": "2026-04-01T06:04:30.916Z",
  "token_count": 5
}

Retrieve pending messages for a registered node. Set ack=true to mark as delivered, ack=false to peek without consuming.

{
  "node_id": "b9e086ffd9a9af44",
  "count": 1,
  "messages": [
    {
      "message_id": "52190bd5-...",
      "sender_id": "9f86d081884c7d65",
      "token_ids": [1, 42, 100, 2000, 500000],
      "intent": "μ",
      "session_id": "conversation-001",
      "encrypted": false,
      "timestamp": "2026-04-01T06:04:30.916Z"
    }
  ]
}

List recent messages across the swarm (admin/debug). Returns message metadata without full token payloads.

Health check endpoint. Returns service status, version, swarm size, and dictionary metrics.