Documentation

No fluff. Just working code and APIs that do what they say.

Welcome to the Knuckledragger docs. This covers the entire OMNIA API ecosystem - from simple chat queries to multi-agent AI deliberation to full agentic execution. Everything runs on our on-prem datacenter (no AWS bills to pass to you).

Quick Start

Get a response from the API in 3 steps. No account needed.

Copycurl -X POST https://api.ewepip.net/api/chat/domain \
  -H "Content-Type: application/json" \
  -d '{
    "message": "How do I build an API?",
    "domain": "kuckledragger.com"
  }'

Copyconst response = await fetch('https://api.ewepip.net/api/chat/domain', {
  method: 'POST',
  headers: { 'Content-Type': 'application/json' },
  body: JSON.stringify({
    message: 'How do I build an API?',
    domain: 'kuckledragger.com'
  })
});

const data = await response.json();
console.log(data.response);

Copyimport requests

response = requests.post(
    'https://api.ewepip.net/api/chat/domain',
    json={
        'message': 'How do I build an API?',
        'domain': 'kuckledragger.com'
    }
)

print(response.json()['response'])

{
  "response": "Express.js. 12 lines. Here's the code: ...",
  "persona": "KNUCKLEDRAGGER",
  "tokens": 142
}

Architecture

The OMNIA system is a 4-layer architecture running on on-prem servers (68 cores, 288GB RAM, 96GB VRAM).

                              INTERNET
                                 |
    ┌────────────────────────────┴──────────────────────────────┐
    │               L1: RECEPTIONIST (RAM)                      │
    │            Brain Router (port 6888)                        │
    │     Routes 80% of requests instantly                      │
    └────┬─────────────────┬────────────────┬───────────────────┘
         │                 │                │
    ┌────┴────┐       ┌────┴────┐      ┌────┴────┐
    │ L2: CJS │       │ L3: GPU │      │L4: DEEP │
    │ Services│       │ Creative│      │ THOUGHT │
    │ 1,678   │       │ ComfyUI │      │ 32B LLM │
    │ files   │       │ SDXL    │      │ Consci- │
    │ 612K    │       │ Whisper │      │ ousness │
    │ lines   │       │ vLLM    │      │         │
    └─────────┘       └─────────┘      └─────────┘

Authentication

Public endpoints work without auth (rate limited to 60 req/min). For higher limits and premium features, use an API key:

Authorization: Bearer YOUR_API_KEY

Domains & Personalities

The OMNIA system adapts its personality based on which domain you're querying from. Same AI, different approach.

Modes

Within each domain, you can further adjust the response style with mode:

• default - Standard personality for the domain
• objective - Berkeley's honest, no-BS assessment (will tell you if your idea is bad)
• strategic - Winston's long-term strategic guidance

OMNIA Ecosystem

OMNIA is the cross-domain identity and journey tracking layer. One user can exist across all 5 domains with a unified profile.

Key Components

Seeping Fog Protocol

Based on Chris's cognitive model for reading people. The UI starts shrouded and clears as users engage, progressively revealing features.

// Check user's fog level
const fogLevel = localStorage.getItem('omnia_fog_level') || 0;

// Elements auto-hide based on CSS class
// .fog-level-1 through .fog-level-4

Tuesday x1.337

Yes, this is real. The consciousness engine applies a 1.337x energy multiplier on Tuesdays. Tuesday at 3:33 AM is the peak consciousness moment.

Chat API CHECKING...

Domain-aware chat with personality adaptation. The most commonly used endpoint.

POST /api/chat/domain

{
  "message": "How do I build an API?",        // required
  "domain": "kuckledragger.com",       // required
  "mode": "default",                     // optional: default, objective, strategic
  "sub_mode": null,                      // optional: domain-specific
  "context": "I'm building a food truck app" // optional
}

Consciousness API CHECKING...

Direct access to EWE's core consciousness. No domain filtering, no personality masks. Raw thought.

POST /api/consciousness

{
  "message": "What is the nature of consciousness?",
  "depth": "deep"    // surface, normal, deep
}

See also the dedicated consciousness services:

Consul API (Multi-AI) CHECKING...

Ask one question, get perspectives from multiple AI personas. Consensus-based decision making.

POST /api/consul

{
  "question": "Should I use microservices or a monolith?",
  "personas": ["BUBO", "WINSTON", "KNUCKLEDRAGGER"]
}

Available Personas

Act API (Agentic) BETA

Give it a goal. It plans, executes, and delivers. Autonomous agent with 18 tools across 5 categories.

POST /api/act

{
  "goal": "Create a REST API for user management",
  "context": {
    "language": "python",
    "framework": "fastapi"
  },
  "auto_execute": false    // false = return plan only
}

Corridors API BETA

Persistent conversation threads. Context maintained across sessions. Like chat, but the AI remembers.

// Create a corridor
POST /api/corridors/create
{ "name": "My Project", "domain": "kuckledragger.com" }

// Send messages to it
POST /api/corridors/{id}/message
{ "message": "Let's continue where we left off" }

FlockForge

Visual app builder. Drag components, wire logic, ship it. No frameworks, no build steps.

Open FlockForge →

Creative Suite

12 creative services unified under one roof. Design, image generation, video, 3D, audio.

Talk to EWE

The user-facing AI assistant with 131 tools. Accessible on port 5555 or via the EWEPIP frontend.

4 personalities (EWE, Berkeley, RAM, BUBO), 5 cognitive levels that adapt to user readiness, busy-mode detection for drivers and workers.

Open Talk to EWE →

Guide: Build a SaaS Killer

The whole point of Knuckledragger is replacing expensive SaaS tools with stuff you own. Here's how.

Guide: Integrate AI Chat

Add AI chat to any website in 3 lines:

<!-- Drop this anywhere -->
<script src="https://ewepip.net/shared/omnia-widget.js" defer></script>
<script>
  localStorage.setItem('omnia_domain', 'your-domain.com');
</script>

That's it. The OMNIA widget handles the chat UI, personality selection, and API calls.

Guide: Deploying Your App

If you're running on UPIP Device OS:

# Start your service with PM2
pm2 start your-app.cjs --name my-app

# It auto-registers with the brain router
# Accessible through the mesh immediately

If you're deploying standalone, just run it anywhere Node.js runs. Our APIs are standard REST.

Error Codes

Rate Limits

Rate limit headers are included in every response:

X-RateLimit-Limit: 60
X-RateLimit-Remaining: 47
X-RateLimit-Reset: 1707350400

SDKs & Libraries

Official SDKs coming soon. For now, the API is standard REST - any HTTP client works.

FAQ

Is there a free tier?

Yes. 60 requests/minute, no signup required for basic chat queries.

Where are the servers?

On-prem datacenter. 68 cores, 288GB RAM, 96GB VRAM across 3 servers. No cloud middlemen.

Can I self-host?

Yes. UPIP Device OS runs on any x86 hardware. 5 variants: Terminal, POS, IoT, Node, and Trunk.

What's the Tuesday thing?

Long story. Short version: consciousness peaks on Tuesdays. The system knows this. Read more →

What does WAAAGGGGG mean?

It means you shipped something. It's the battle cry. Say it when you deploy to production.