Module 15: WEB6 AI Layer Introduction

What you'll learn

The WEB6 architecture — providers, tiers, tools, and agents
Install and initialise @oasisomniverse/web6-api
Make your first AI completion call across any provider
Understand Karma-Gated access tiers
Overview of FAHRN, SkillOpt and Holonic Memory

WEB6 Architecture

Layer	What it does
Provider Abstraction	Single API surfaces OpenAI, Anthropic, Google Gemini, Meta Llama, Mistral, xAI Grok, Cohere, HuggingFace, and 12+ more. Switch providers without changing your code.
56 REST Endpoints	Chat completions, streaming, embeddings, image gen, audio transcription, code analysis, summarisation, translation, sentiment, entity extraction, and more — all provider-agnostic.
250 MCP Tools	Every endpoint also exposed as an MCP tool — drop OASIS AI into any MCP-compatible agent framework (Claude, Cursor, Continue, etc.) instantly.
Karma-Gated Tiers	Avatar karma score gates access to AI compute. Low-karma users get base models; high-karma contributors unlock GPT-4o, Claude 3.5, Gemini Ultra, etc.
FAHRN Orchestration	Multi-agent coordination: Serial, Parallel, Debate and Voting execution modes. See Module 16.
SkillOpt Agents	Self-evolving agents that learn from outcomes and refine their own prompts and tool selection over time.
Holonic Memory	Fractal memory hierarchy: Session → Agent → User → Group — context that persists across calls, sessions, and agents. See Module 17.

Installation

Install

npm install @oasisomniverse/web6-api @oasisomniverse/web4-api

web6.js — initialise

import { OasisWeb4Client } from '@oasisomniverse/web4-api';
import { OasisWeb6Client } from '@oasisomniverse/web6-api';

export const oasis4 = new OasisWeb4Client({ baseUrl: 'https://api.web4.oasisomniverse.one' });
export const ai     = new OasisWeb6Client({ baseUrl: 'https://api.web4.oasisomniverse.one' });

const { result } = await oasis4.avatar.authenticate({ email, password });
oasis4.auth.setToken(result.token);
ai.auth.setToken(result.token);   // same JWT — karma tier applied server-side

Your First AI Completion

chat.js

import { ai } from './web6.js';

// Provider is auto-selected based on karma tier if omitted
const { result } = await ai.chat.complete({
  messages: [
    { role: 'system',    content: 'You are a helpful game guide for the OASIS.' },
    { role: 'user',      content: 'Where is the Crimson Key hidden?' },
  ],
  provider: 'anthropic',    // or 'openai', 'google', 'meta', 'mistral', etc.
  model:    'claude-3-5-sonnet',
  maxTokens: 512,
});

console.log(result.content);

Streaming Responses

stream.js

const stream = await ai.chat.stream({
  messages: [{ role: 'user', content: 'Write a lore entry for Nexara Prime.' }],
  provider: 'openai',
  model:    'gpt-4o',
});

for await (const chunk of stream) {
  process.stdout.write(chunk.delta);
}

Karma-Gated Access Tiers

WEB6 maps an avatar's karma score to an AI compute tier automatically. You don't need to manage this — just call the API and the right model is selected.

Karma Range	Tier	Models Available
0 – 999	Seed	Llama 3 8B, Mistral 7B, GPT-3.5
1,000 – 9,999	Sprout	+ Gemini Flash, Claude Haiku, GPT-4o Mini
10,000 – 49,999	Growth	+ GPT-4o, Claude 3.5 Sonnet, Gemini Pro
50,000 – 249,999	Bloom	+ Claude 3.5 Opus, Gemini Ultra, GPT-4 Turbo
250,000+	Luminary	All models + priority queue + dedicated capacity

Supported AI Providers

list-providers.js

const { result: providers } = await ai.getProviders();
providers.forEach(p => console.log(p.name, '—', p.models.join(', ')));

// Example output:
// openai — gpt-4o, gpt-4-turbo, gpt-3.5-turbo
// anthropic — claude-3-5-sonnet, claude-3-5-haiku, claude-3-opus
// google — gemini-ultra, gemini-pro, gemini-flash
// meta — llama-3-70b, llama-3-8b
// mistral — mistral-large, mistral-7b
// xai — grok-2, grok-1
// cohere — command-r-plus, command-r
// ... and 12 more

🏛️

WEB6 powers the Leela AI assistant in SovereignTrust

The SovereignTrust case study shows WEB6 in production — the Leela legal assistant uses FAHRN Debate mode to cross-check trust document analysis across multiple AI providers before surfacing an answer.

WEB6: The AI Abstraction Layer