WEB6 · THE INTELLIGENCE LAYER

OASIS WEB6

WEB6 v2.0 MAJOR ARCHITECTURE UPGRADE
JULY 2026
ONE API. EVERY AI.

The unified AI abstraction and aggregation layer — now with self-evolving agent skills (SkillOpt), in-process ML inference (ML.NET), DID/Verifiable Credential identity, karma-gated AI and full ACP/ANP/gRPC/GraphQL protocol support. Built on WEB4 and WEB5, routing intelligently across 20+ providers.

EXPLORE THE API → GRAPHQL → NPM PACKAGE → READ THE WHITEPAPER →
WEB6 API OPENAI GPT-4o ANTHROPIC CLAUDE GEMINI GOOGLE OLLAMA LOCAL MISTRAL MIXTRAL LLAMA META GROQ LLAMA COHERE COMMAND HF MODELS XAI GROK MCP · A2A PROTOCOLS AGENTS ORCHESTRATION
MAJOR RELEASE · JULY 2026

What's New in WEB6 v2.0

A complete 25-point architecture upgrade — from self-evolving AI skills and in-process ML to DID identity, multi-protocol orchestration and karma-gated access. This is the platform investors and enterprise teams have been waiting for.

NEW
SELF-EVOLVING AI SKILLS

SkillOpt (Microsoft Research, arXiv:2605.23904) — each FAHRN agent evolves its own best_skill.md procedure through a rollout → reflect → edit → gate loop. +23.5% avg gain demonstrated. No weight updates — skills transfer across models and harnesses. Paired with ML.NET in-process AutoML at microsecond latency.

NEW
MULTI-PROTOCOL ORCHESTRATION

ACP (BeeAI/IBM), ANP (DID-based discovery), gRPC bidirectional streaming, GraphQL subscriptions, and AsyncAPI/Kafka event mesh — all behind a single unified endpoint. FAHRN agents can now be discovered, invoked and coordinated across any agent runtime on the internet.

NEW
DID / VERIFIABLE CREDENTIALS

W3C DID + Verifiable Credential support (did:key, did:web, did:ethr, did:ion). AI agents act on behalf of avatars via cryptographically-signed VC grants. Every FAHRN dispatch is attributable to a provable, consent-governed identity — critical for enterprise compliance.

NEW
KARMA-GATED AI ACCESS

Avatar karma tiers gate model access and monthly token budgets — Bronze through Diamond. Karma earns through contributions, OAPPs, community and verified credentials. Creates a self-sustaining incentive loop where high-quality participation unlocks more powerful AI capabilities.

NEW
HOLONIC MEMORY UPGRADES

External memory provider integration (Mem0, Zep, Letta, LangMem, Graphiti, Redis Vector), upward memory propagation API, and WebSocket bidirectional sessions with persistent server-side state in Holonic BRAID — turning every conversation into a living memory artifact.

NEW
FULL OBSERVABILITY

OpenTelemetry traces and Prometheus metrics on every FAHRN dispatch, memory operation and provider call. Grafana dashboards, structured JSON logging, cost-per-dispatch analytics and real-time alerting — production-grade telemetry from day one.

SEE THE FULL ROADMAP → READ THE v2.0 WHITEPAPER →
THE CONCEPT

WHAT IS WEB6?

The AI landscape is fragmented — dozens of providers, incompatible APIs, competing agent protocols and orchestration frameworks. WEB6 is the abstraction layer that unifies them all, so you build once and deploy everywhere.

YOUR APPLICATION
ONE UNIFIED INTERFACE
↓ CALLS
WEB6 OASIS AI API  ·  api.web6.oasisomniverse.one
ABSTRACTION LAYER
↓ ROUTES TO
OpenAI  ·  Anthropic  ·  Google Gemini  ·  Meta Llama  ·  xAI Grok
LLM PROVIDERS
Mistral  ·  Groq  ·  Cohere  ·  Ollama  ·  Hugging Face  ·  +more
OPEN-SOURCE & LOCAL
LangChain  ·  AutoGen  ·  CrewAI  ·  MCP  ·  A2A Protocol  ·  Semantic Kernel
AGENT FRAMEWORKS
AI PROVIDERS

EVERY MODEL. ONE ENDPOINT.

WEB6 normalises the request and response interface for every major AI provider and open-source model runtime. Switch providers with a single config change — no refactoring required.

🤖
OPENAI
GPT-5 · GPT-4o · o3 · DALL·E · Whisper
ANTHROPIC
Claude Opus · Sonnet · Haiku
💎
GOOGLE GEMINI
Gemini 2.5 · Flash · Pro · Veo
🦙
META LLAMA
Llama 3.x · Code Llama · Scout
🌀
MISTRAL
Mixtral · Mistral Large · Codestral
🔥
GROQ
Ultra-fast LPU inference
🏠
OLLAMA
Local · Private · Self-hosted
🤗
HUGGING FACE
Open-source · Fine-tuned models
🧠
COHERE
Command R+ · Embed · Rerank
XAI
Grok 3 · Grok Vision
🐋
DEEPSEEK
R1 · V3 · Coder
☁️
AWS BEDROCK
Titan · Nova · Jurassic
🔷
AZURE OPENAI
Enterprise GPT · DALL·E · Vision
🎨
STABILITY AI
SDXL · Stable Diffusion · SVD
🌐
OPENSERV
One SERV key → GPT-5 · Claude · Gemini · Grok · Qwen · DeepSeek
🔴
REPLICATE
Open-source model marketplace · Image · Audio · Video
🧬
BITTENSOR
Decentralised inference · Community owned · Coming soon
🌿
GAIANET
Decentralised community nodes · Coming soon
⚙️
GENSYN
Decentralised GPU compute · Verifiable inference · Coming soon
☁️
AKASH
Decentralised cloud · Self-hosted models · Coming soon
CAPABILITIES

WHAT WEB6 GIVES YOU

More than a proxy — WEB6 is an intelligent orchestration and normalisation layer sitting on top of both WEB4 and WEB5. Every AI request is grounded with real-world identity, cross-reality app data and a living holon graph — making it the most contextually rich AI API ever built.

🔀
UNIFIED API
One REST endpoint for every provider. Same request shape, same response schema. Swap models with a single parameter change — no client-side refactoring needed.
🧭
INTELLIGENT ROUTING
Route requests automatically to the best model for the task — by capability, latency, cost or availability. Define routing rules declaratively, or let WEB6 decide.
🤝
AGENT ORCHESTRATION
Coordinate multi-agent workflows across heterogeneous AI runtimes. Chain agents, pass context, monitor progress and handle failures — through one unified interface.
📡
PROTOCOL NORMALISATION
Abstracts over MCP, A2A, LangChain, AutoGen, CrewAI and Semantic Kernel. Write once against the WEB6 protocol and reach every agent framework.
🛡️
FAILOVER & LOAD BALANCING
Automatic fallback to secondary providers when a model is unavailable or rate-limited. Built-in retry logic, circuit breaking and cost-aware load distribution.
🌐
WEB4 UNIFICATION + WEB5 XP
WEB6 sits on top of both WEB4 (unification, karma, COSMIC ORM — 40+ storage providers) and WEB5 (STAR ODK, STARNET holon graph, all cross-world OAPP data). Every AI call is automatically grounded with the world's richest cross-reality dataset.
💰
COST OPTIMISATION
Route cheaper tasks to smaller models automatically. Track spend per provider, per user and per OASIS avatar. Set budgets and caps — WEB6 enforces them.
📊
OBSERVABILITY
Unified logs, latency metrics and token usage across every provider. Debug multi-agent workflows with end-to-end trace IDs that span the full OASIS stack.
🔑
SINGLE AUTH
One OASIS avatar key authenticates against every provider. No per-provider API key management. Permissions and quotas are governed at the WEB4 UNIFICATION LAYER.
STREAMING SSE
Real-time token-by-token streaming via Server-Sent Events across every provider — normalised to a single stream format regardless of the underlying model's native streaming protocol.
🧮
EMBEDDINGS & VECTOR SEARCH
Unified embeddings endpoint across OpenAI, Cohere, Hugging Face and local models. Powers semantic caching (95% similarity threshold), RAG pipelines and holonic memory search — all through the same normalised API.
🛠️
TOOL / FUNCTION CALLING
Normalised tool-use interface across every provider that supports function calling. Define tools once — WEB6 translates the schema and result format to each provider's native format automatically.
🔁
WEBSOCKET SESSIONS
Persistent bidirectional agent sessions via WebSocket. Server-side conversation state, mid-session tool results, user interrupts and multi-turn FAHRN loops — all backed by Holonic Memory so sessions survive reconnects.
🏅
KARMA-GATED AI
Avatar karma score gates AI access tier. Low-karma users get cost-efficient models; high-karma users unlock the best models with optimised routing. Earn karma in the OurWorld game → unlock GPT-5 and Claude Opus. The world's first karma-gated AI API.
🪪
DID / VERIFIABLE CREDENTIALS
W3C-standard decentralised identity auth (did:key, did:web, did:ethr). AI agents can act verifiably on behalf of avatars via Verifiable Credential capability grants — revocable, on-chain provenance, no centralised auth server needed.
🧠
EXTERNAL MEMORY PROVIDERS
Plug in Mem0, Zep, Letta/MemGPT, LangMem, Graphiti or Redis Vector Memory alongside the fractal Holonic Memory hierarchy. Long-term agent memory that works with whatever memory layer you already use.
SEMANTIC CACHING
Embedding-based prompt cache: if an incoming request is ≥95% similar to a cached prompt, the cached response is returned instantly at zero provider cost. Dramatic cost and latency reductions for repeated or near-identical queries.
🔬
IN-PROCESS ML INFERENCE
Powered by ML.NET — Microsoft's open-source C# ML framework (used in Power BI, Defender, Bing). Runs sentiment analysis, anomaly detection, task classification and recommendation models in-process inside the OASIS API — zero additional API calls, microsecond latency, and full AutoML pipeline for training on OASIS usage data.
⚖️
DEBATE & VOTING MODES
Two additional FAHRN dispatch modes for high-stakes decisions: Debate pits agents against each other adversarially with NLI contradiction checking; Voting runs independent agents and selects by consensus. The only AI API with built-in multi-model adversarial reasoning.
🔮
STARNET CONTEXT MANAGER
One call assembles a rich avatar context block pulling Web4 identity, Web5 XP, karma tier, active holons and past AI session patterns in parallel. Every FAHRN dispatch is grounded in the avatar's full cross-reality profile — zero extra round trips from the caller.
📡
MULTI-PROTOCOL ORCHESTRATION
ACP (BeeAI/IBM), ANP (DID-based agent discovery), gRPC bidirectional streaming, GraphQL subscriptions and AsyncAPI/Kafka event mesh — all behind a single endpoint. FAHRN agents can be discovered and coordinated across any agent runtime on the internet, not just OASIS.
♾️
SELF-REGISTRATION AS ORCHESTRATOR
On startup the WEB6 API registers itself as its own MCP orchestrator — exposing every FAHRN agent, memory provider and protocol adapter as a callable MCP tool. Claude Code, IDE extensions and third-party agents discover the full capability surface automatically. Zero manual wiring.
🌍
COLLECTIVE INTELLIGENCE
Aggregate routing performance, skill evolution signals and anonymised task outcomes propagate up the Holonic BRAID hierarchy through membrane rules. A city-level holon develops awareness of which agents solve which problems best — shared intelligence that benefits every user without exposing any individual session.
🛡️
ANTI-FRAGILE ROUTING
Every agent stall, timeout and failure updates composite routing scores via EMA — the system becomes more accurate precisely because of the failures it encounters. Circuit breaker recovery, score decay, zero-downtime provider swap and ML anomaly detection combine to make WEB6 more reliable under adversity, not less.
THE ENDPOINT

ONE ENDPOINT. ALL MODELS.

Authenticate with your OASIS avatar key and start making requests immediately — no per-provider SDK setup, no credential juggling.

api.web6.oasisomniverse.one
// Route a completion to whichever model fits best
POST https://api.web6.oasisomniverse.one/v1/complete
Authorization: Bearer <your-oasis-avatar-key>

{
  "provider": "auto",          // "auto" | "openai" | "anthropic" | "gemini" | "groq" | "mistral" | ...
  "model":    "auto",          // or a specific model ID
  "messages": [
    { "role": "user", "content": "Hello from the OASIS" }
  ],
  "routing": {
    "priority": "cost",         // "cost" | "latency" | "quality"
    "fallback": true
  },
  "avatarId": "your-avatar-id"     // optional — injects WEB4 identity/karma context
}
// JavaScript / Node.js — install the npm package
npm install @oasisomniverse/web6-api

import { Web6Client } from '@oasisomniverse/web6-api';

const web6 = new Web6Client({ baseUrl: 'https://api.web6.oasisomniverse.one' });
web6.setToken('<your-oasis-jwt>');

const { result } = await web6.completion.complete({
  Provider: 'auto',
  Model:    'auto',
  Messages: [{ Role: 'user', Content: 'Hello from the OASIS' }],
  UseFAHRN:        true,   // optional: FAHRN reasoning pipeline
  UseHolonicBraid: true,   // optional: shared graph injection
  Routing: { Fallback: true },
});
console.log(result.Content);
// Dispatch a problem straight to FAHRN - the controller agent scores agents, picks a
// dispatch mode, runs loop detection and returns the assembled Mermaid execution plan
POST https://api.web6.oasisomniverse.one/v1/reasoning-network/dispatch
Authorization: Bearer <your-oasis-avatar-key>

{
  "problem":  "Plan a zero-downtime migration of a 50M row table",
  "taskType": "architecture",
  "mode":     "Parallel"   // "Serial" | "Parallel" | "Decomposed"
}
DEEP INTEGRATION

POWERED BY WEB4 + WEB5

WEB6 doesn't just abstract AI providers — it sits on top of the entire OASIS data stack. WEB4 and WEB5 feed their world data directly into the AI layer, making every request aware of identity, reputation, cross-world state and a living ecosystem of apps and holons.

WEB4 — IDENTITY & DATA LAYER

The universal identity, reputation and storage unification layer. Every AI request in WEB6 can be grounded with your avatar's full history, karma score and COSMIC ORM data across 40+ storage providers — giving AI models context that no other platform can offer.

AVATAR IDENTITY KARMA SCORES COSMIC ORM CROSS-WORLD MEMORY 40+ PROVIDERS
EXPLORE WEB4 →

WEB5 — APP & HOLON LAYER

STAR ODK and STARNET publish every app, asset and interaction as a versioned holon in a living graph. WEB6 pulls this entire cross-world application data stream — all OAPPs, user interactions and holon state — directly into the AI layer, making it extraordinarily context-aware.

STAR ODK STARNET GRAPH HOLON DATA OAPP ECOSYSTEM CROSS-WORLD STATE
EXPLORE WEB5 →
THE RESULT

WEB6 is the only AI abstraction layer that feeds real-world identity and cross-chain data (WEB4) and a live ecosystem of cross-world apps and holons (WEB5) directly into every AI model call — making it the most powerful and contextually rich AI API ever built.

BRAID · FRACTAL HOLONIC SHARED MEMORY · CROSS-CHAIN PERSISTENCE

HOLONIC BRAID

OASIS extends the OpenSERV BRAID framework (Amçalar & Cinar, arXiv:2512.15959) with a fractal holonic shared graph library — delivering 74× PPD gains on GSM-Hard that scale indefinitely across agents and sessions. One shared library. Every reasoning graph ever generated — available to everyone, forever.

TWO-STAGE BRAID PROTOCOL
GENERATOR
High-tier model (gpt-4.1) creates a Mermaid reasoning graph for task type τ — once per task type, stored as a holon in the shared library.
SOLVER
Low-tier model (gpt-5-nano-minimal) executes the graph against each task instance — at a fraction of the cost, with equal or superior accuracy.
74×
PPD on GSM-Hard
30×
PPD on procedural tasks
94%→98%
accuracy on GSM-Hard
🌍 EARTH HOLON
Universal collective intelligence
🌐 CONTINENT → COUNTRY → COUNTY → CITY → DISTRICT → NEIGHBOURHOOD
Geographic holon chain — memory propagates upward through membrane rules
👥 ORG / GROUP HOLONS
Friends · Family · Colleagues · Clubs · Communities (user can belong to many)
👤 USER HOLON
All agents for this user — unified identity & reputation
🤖 AGENT HOLON
All sessions for this agent, for this user
💬 SESSION HOLON
One conversation — memory stored as a holon
📚
SHARED GRAPH LIBRARY
Reasoning graphs generated by any agent are stored as holons in a shared library. Every future agent doing the same task type retrieves the graph at zero generation cost — PPD gains compound globally rather than resetting per agent.
⛓️
CROSS-CHAIN PERSISTENCE
Graph holons are replicated across MongoDB (fast access), Solana (immutable provenance) and IPFS (decentralised permanence) via the COSMIC ORM. Each holon carries a ProviderUniqueStorageKey per backend — same graph, any provider.
🧬
FRACTAL MEMORY HIERARCHY
Every AI session generates a session holon. That holon is a child of the agent holon, which is a child of the user holon — memory and intelligence compound upward through every level of the hierarchy, from session to Earth.
🔬
MEMBRANE RULES
At each boundary between holons, user-configurable membrane rules define which parts of the memory are persisted, propagated or kept private. Per-field granularity — nothing leaks without explicit permission.
🌍
COLLECTIVE INTELLIGENCE
Session → Agent → User → Group → Neighbourhood → City → Country → Earth. Collective intelligence flows upward from every individual interaction to the planetary level — genuinely bottom-up, consent-governed emergence.
♾️
A PATH TO AGI
BRAID's PPD gains + holonic shared memory + geographic propagation = a system that improves with every interaction, at every scale. Modelled after nature, unity consciousness and the infinite intelligence of the universe.
READ THE FULL WHITEPAPER → ORIGINAL LITEPAPER →
META-ORCHESTRATION INTELLIGENCE LAYER · COMPOSITE SCORING · CONTINUOUS LEARNING · ANTI-FRAGILE

FAHRN

FRACTAL ADAPTIVE HOLONIC REASONING NETWORK

FAHRN is the self-optimising reasoning mesh built on top of Holonic BRAID. A controller agent dispatches work to scored agents in serial, parallel or decomposed mode — using a composite score of category performance, speed, cost and loop-detection stability — and learns from every outcome via EMA score updates stored in the holonic memory hierarchy.

CONTROLLER AGENT — DISPATCH LOGIC
🤖
GPT-5
Reasoning ★★★★★
Speed ★★★★☆
Code ★★★★☆
CLAUDE
Analysis ★★★★★
Speed ★★★★☆
Writing ★★★★★
GROK
Real-time ★★★★★
Speed ★★★★★
Search ★★★★★
+
MORE AGENTS
Each scored per
category & speed
continuously
COMPOSITE ROUTING SCORE
CompositeScore =
  (Wcat × CategoryScore)
  + (Wspeed × SpeedScore)
  + (Wcost × CostScore)
  − (Wpenalty × FailureRate)
  − (LoopPenalty if recently stalled)
Weights are configurable per mode — Serial mode elevates Wcost; Parallel mode favours Wcat; Decomposed mode raises Wspeed to minimise wall-clock time. Agent metadata includes: category_scores, speed_score, cost_score, loop_detection_score, failure_rate — all updated continuously from real outcomes.
MODE 1 — SERIAL
COST OPTIMISED

The controller dispatches the problem to the highest-scoring agent for that category. If it stalls, the next best agent takes over. One agent works at a time — lowest token cost, still intelligent routing.

Lowest cost
Automatic fallback on timeout
Generates Mermaid execution plan
MODE 2 — PARALLEL
ACCURACY OPTIMISED

All agents receive the same problem simultaneously. Each produces a Mermaid execution diagram. The controller compares plans — selecting the best, or merging complementary diagrams into a superior combined plan.

Highest accuracy
Plan comparison & merging
Best-of-all-worlds execution
MODE 3 — DECOMPOSED
COMPLEX PROBLEMS

Large, complex problems are broken into sub-problems. Each is routed to the agent best matched to its category and speed profile. The resulting sub-diagrams are then combined by the controller into one unified execution plan.

Handles unlimited complexity
Best-fit agent per sub-problem
Merged composite diagram
🎛️
CONTROLLER AGENT
A meta-level orchestrator sits above the reasoning network. It classifies problems, selects dispatch mode, ranks agents by composite score, monitors for timeout/stall, promotes fallback agents and assembles the final execution plan.
📊
COMPOSITE SCORING
Every agent carries a full metadata object: category scores (mathematics, legal, architecture, code, writing, real-time…), speed_score, cost_score, loop_detection_score, failure_rate. The composite score formula blends all dimensions with user-configurable weights per mode.
🔁
LOOP DETECTION
Four detection mechanisms run in parallel: output similarity hashing (repeated reasoning patterns), self-contradictory step detection, token budget monitoring, and circular DAG detection in the Mermaid output. Any trigger suspends the agent and escalates immediately.
📐
MERMAID EXECUTION PLANS
Each agent produces a structured Mermaid diagram of its execution plan — machine-comparable, mergeable and human-readable. In parallel mode plans are compared and merged. In decomposed mode sub-diagrams are stitched into one coherent whole.
📈
CONTINUOUS LEARNING
After every task, all agent scores update via EMA (Exponential Moving Average): category scores, speed, failure rate, loop stability. Updates are stored as holons in the Holonic BRAID hierarchy — the system routes better with every single interaction, forever.
🛡️
ANTI-FRAGILE BY DESIGN
Every stall, timeout and failed plan updates agent metadata and improves future routing. No single-model bias — performance decides flow, not reputation. No provider lock-in — the system naturally migrates workloads toward consistently better agents. Each agent is itself a holon.
📄
SELF-EVOLVING SKILL DOCUMENTS
Powered by Microsoft Research's SkillOpt (arXiv:2605.23904). Each FAHRN agent carries a portable best_skill.md artifact — a natural-language procedure document that evolves through a rollout → reflect → edit → gate loop, without touching model weights. Demonstrated +23.5% average gain across benchmarks. Skills transfer across models, harnesses and task categories — stored as holons in Holonic BRAID.
🗺️
MERMAID EXECUTION PLANS
For complex multi-step problems the controller generates a Mermaid flowchart execution plan before dispatching — decomposing the problem into subtasks, assigning the optimal agent to each node, and assembling the final answer from the graph traversal. Plans are stored as holons so future sessions can query and reuse them.
🔮
KARMA-GATED AI ACCESS
Avatar karma tiers (Bronze → Silver → Gold → Platinum → Diamond) gate model access and monthly token budgets. Karma is earned through contributions, OAPP publishing, community participation and verified credentials — creating a self-sustaining incentive loop where high-quality participation unlocks progressively more powerful AI capabilities.
FULL SYSTEM STACK
READ THE FULL WHITEPAPER → ANTI-FRAGILITY & STACK →
MODEL CONTEXT PROTOCOL

OASIS MCP

Connect any MCP-compatible AI agent or tool directly to the full OASIS platform — avatar unification layer, karma, COSMIC ORM data and the entire WEB6 AI layer — through a single standardised Model Context Protocol interface.

🔌
PLUG IN ANY AGENT
Any MCP-compatible agent — Claude, Cursor, Continue, or your own — can call OASIS APIs natively. No custom integration code required.
👤
AVATAR CONTEXT
Agents gain full access to OASIS avatar identity, karma scores and cross-world reputation — grounding every AI response in real user context.
🧬
COSMIC ORM DATA
Read and write to the universal COSMIC ORM layer — 40+ providers, unified schema — directly from your AI agent workflows via MCP tool calls.
🤖
WEB6 AI ROUTING
MCP tool calls route through the full WEB6 abstraction layer — intelligent model selection, failover and cost optimisation built in.
🔑
SINGLE AUTH
One OASIS avatar key authenticates the MCP server against the entire platform. No per-provider credentials to manage or rotate.
📡
STANDARD PROTOCOL
Built on Anthropic's open Model Context Protocol standard — interoperable with every MCP host and any agent framework that supports it.
HTTP MCP TRANSPORT — CLOUD & REMOTE ACCESS

The MCP server also runs as an HTTP endpoint built into the WEB6 REST API — for cloud-hosted agents, claude.ai connectors, and any client that cannot use stdio:

GET  /mcp                     — SSE stream (HTTP MCP transport)
POST /mcp                     — MCP JSON-RPC tool calls
GET  /.well-known/mcp.json    — auto-discovery document
GET  /.well-known/agent.json  — A2A agent card

These endpoints are not shown in the Swagger UI — they use SSE/streaming transports the interactive explorer cannot represent. Point your MCP client at the discovery URL and it configures itself automatically. Full details in the MCP Server README.

GET OASIS MCP →
AI-NATIVE DEVELOPMENT

OASIS IDE

A desktop coding environment built around the same WEB6 AI layer — agentic coding assistants (Claude, OpenServ and every model in the SERV catalog), MCP tool access and full OASIS/STAR project awareness, in one editor.

🤖
AGENTIC CODING
Run Claude Sonnet 4.6 or any OpenServ-catalog model as a tool-using coding agent directly against your open workspace — read, write, search and run commands with explicit confirmation.
🔌
MCP-AWARE
Connects to the OASIS MCP server out of the box, giving every agent native access to avatar identity, karma and COSMIC ORM data while you code.
🧬
WEB6 ROUTING
Agent mode runs through the same multi-provider WEB6 abstraction layer — switch between Claude and the full SERV catalog (OpenAI, Gemini, Grok, Qwen, DeepSeek) from one toggle.
🗂️
OASIS/STAR AWARE
Built for working inside OASIS and STAR ODK workspaces — file search, terminal and project context tuned for this codebase's structure.
🧠
FAHRN-POWERED PLANNING
Invoke the full FAHRN reasoning network from inside the IDE — decompose complex refactors, architectural changes or migrations into scored Mermaid execution plans before a single line of code is touched.
💾
PERSISTENT PROJECT MEMORY
Every coding session is stored as a holon in the Holonic BRAID hierarchy. Decisions, patterns and architectural context persist across restarts — the IDE gets smarter about your codebase over time.
GET OASIS IDE →
THE OPPORTUNITY

WHY WEB6. WHY NOW.

The AI infrastructure market is fragmenting at exactly the moment enterprises need consolidation. WEB6 is the only platform that unifies the entire stack — from models to memory to identity — and already has 8 years of production infrastructure underneath it.

$1.8T
AI MARKET BY 2030
Every enterprise building on AI needs a provider-neutral abstraction layer. WEB6 is that layer — already built, already running.
15+
AI PROVIDERS, ONE API
No other platform normalises OpenAI, Anthropic, Gemini, Grok, Groq, Mistral, DeepSeek, Ollama, AWS Bedrock, Azure, OpenServ, Replicate and Hugging Face behind a single unified endpoint.
8 YRS
PRODUCTION INFRASTRUCTURE
WEB4 and WEB5 underpin WEB6 — 40+ storage providers, real-time holon graph, cross-chain identity and karma, all production-hardened over 8 years of active development.
74×
PPD GAIN — HOLONIC BRAID
Peer-reviewed 74× performance-per-dollar improvement on GSM-Hard benchmarks. No other commercial AI API ships a shared reasoning graph library with provable compound gains.
UNIQUE DATA MOAT
Every AI call is grounded in real-world avatar identity, karma, cross-world quests and a live holon graph. No competitor can replicate this data foundation without rebuilding WEB4 and WEB5 from scratch.
WEB4→10
COMPLETE STACK OWNERSHIP
OASIS is the only platform that owns the entire stack from data unification (WEB4) through AI abstraction (WEB6), human-AI symbiosis (WEB7), galactic mesh (WEB8), singularity monitoring (WEB9) to root identity (WEB10).
THE MOAT

WEB6 is not a wrapper around OpenAI. It is an intelligence operating system — with its own reasoning network (FAHRN), its own shared memory layer (Holonic BRAID), its own identity system (WEB4 karma + DID/VC), and a data moat built over 8 years that no new entrant can replicate. The protocol-neutral architecture means it gets stronger with every new AI model that launches — each new provider is another node in the network.

INVESTOR GUIDE → FOUNDERS NFT →
THE PLAN

ROADMAP

WEB6 v2.0 is fully operational — all 25 improvements across Phases 1–3 are shipped as of July 2026. Phase 4 is the long-horizon AGI pathway.

✅ COMPLETE
PHASE 1 — FOUNDATIONS
Q1–Q2 2026
15+ AI providers unified under one REST endpoint
FAHRN — controller agent, composite scoring, Serial / Parallel / Decomposed modes, EMA continuous learning
Holonic BRAID — shared reasoning graph library, fractal memory hierarchy, cross-chain persistence
OASIS MCP server — 250 tools across WEB4–WEB10, in-process, zero HTTP overhead
OrchestratorManager — MCP, A2A, LangChain, AutoGen, CrewAI, Semantic Kernel behind one interface
OpenServ SERV gateway — one API key reaches GPT-5, Claude, Gemini, Grok, Qwen, DeepSeek
Image generation (StabilityAI + OpenAI), JWT auth, OASIS DNA config
WEB7–WEB10 APIs live — symbiosis, galactic mesh, singularity, source
✅ COMPLETE
PHASE 2 — DEVELOPER EXPERIENCE
COMPLETE · JULY 2026
MCP HTTP / Streamable HTTP transport — expose OASIS MCP to any cloud host or claude.ai connector
Streaming SSE — real-time token streaming normalised across all 15+ providers
Tool / function calling — unified schema, provider-native translation
Embeddings endpoint — OpenAI, Cohere, HuggingFace + local; powers RAG and semantic cache
POST /v1/fahrn/solve — single hero endpoint: problem in, reasoned answer + Mermaid plan + telemetry out
A2A proper protocol — Agent Card, task lifecycle, SSE stream
WebSocket bidirectional sessions — persistent agent state, mid-session tool results, FAHRN interrupt
Karma-gated AI access — earn karma → unlock better models and higher monthly budgets
Per-avatar cost metering, monthly budgets and quota enforcement
Auto-generated SDK — @oasisomniverse/web6-api npm + oasis-web6 Python package
✅ COMPLETE
PHASE 3 — INTELLIGENCE EXPANSION
COMPLETE · JULY 2026
DID / Verifiable Credentials — W3C decentralised identity auth, VC capability grants, AI acting verifiably on behalf of avatars
ACP (BeeAI/IBM), ANP (Agent Network Protocol), gRPC, GraphQL and AsyncAPI adapters
External memory providers — Mem0, Zep, Letta/MemGPT, LangMem, Graphiti, Redis Vector alongside Holonic BRAID
Semantic caching — embedding-based prompt deduplication at 95%+ similarity; near-zero cost on repeated queries
Enhanced FAHRN loop detection — output hashing, DAG cycle detection, NLI contradiction checking, token budget hard stops
Holonic Memory — multi-hop propagation, semantic search within holons, TTL retention policies
Real-time telemetry SSE stream — per-request traces: provider, latency, FAHRN path, BRAID hit/miss, cost
Decentralised AI — Bittensor (Corcel gateway), GaiaNet, Replicate as first-class providers
StarnetContextManager — standalone GET /v1/context/avatar/{id} rich context block for any external system
FAHRN Debate & Voting dispatch modes — adversarial multi-agent reasoning for high-stakes decisions
SkillOpt self-evolving agent skills — rollout → reflect → edit → gate loop produces portable best_skill.md holons; demonstrated +23.5% avg gain (Microsoft Research, arXiv:2605.23904)
ML.NET in-process inference — C#-native AutoML, sentiment analysis, anomaly detection, task classification and agent routing recommendations running at microsecond latency inside the OASIS API
🌌 2027
PHASE 4 — THE AGI PATHWAY
2027
Decentralised AI marketplace — open model registry, community-contributed inference, tokenised provider incentives
WEB7 symbiosis deep integration — bio-signal context (EEG/HRV) feeding directly into FAHRN routing decisions
WEB8 galactic mesh routing — AI tasks routed across interplanetary relay nodes with Dijkstra path optimisation
Holonic BRAID at Earth scale — every session in the OASIS ecosystem contributes to a planetary reasoning graph
Collective consciousness layer — group holons surface emergent intelligence from millions of individual agent sessions
WEB9 SingularityAggregationManager — unified health, capability and intelligence score across WEB4–WEB8
FAHRN self-improvement loop — reasoning network that rewrites its own agent scoring weights based on outcomes
AGI milestone tracking via WEB9 — real-time dashboard of collective intelligence metrics across the OASIS omniverse
INVESTOR EVALUATION GUIDE → JOIN AS A FOUNDER →