Vision

B3IQ's role as the AI, proof, and settlement layer for the B3 ecosystem.

B3IQ is a neocloud for AI inference: customers buy or rent-to-own bare-metal GPU servers that B3IQ hosts or ships and operates as managed, manually onboarded nodes. It is also the AI, proof, and settlement layer for the wider B3 ecosystem — B3 apps, agents, and execution systems can call B3IQ as standard HTTP infrastructure, then receive model output, usage metadata, receipts, and settlement evidence. Routing is own-nodes-strict and pricing/payouts are in USD.

The Contract

B3IQ does not need to become an agent marketplace, workflow system, or SDK vendor. The product contract is:

  • OpenAI-compatible model listing and chat completions.
  • B3IQ-native private job, receipt, billing, node-status, and route-readiness APIs.
  • Public-safe protocol evidence for benchmark, settlement, reputation, and disputes.
  • Operator tooling for install, setup, local auth, runtime readiness, and safe remote operations.

Target Architecture

B3IQ Target Role
HTTP requestnormalize + quoteeligible routesigned receiptroots + payoutsmetadata B3 Apps And Agents Standard HTTP clients B3IQ Endpoints OpenAI-compatible + native APIs Policy Layer Routing, billing, privacy, abuse gates Node Supply Local, portable, network beta, future fleet tiers Proof And Receipts Usage, latency, TTFT, benchmark, settlement evidence Settlement Network Registry, registration state, roots, payout proofs

B3 Ecosystem Fit

B3OS rounded circle logo

B3OS B3OS and agent execution layers can call B3IQ as an endpoint-first work provider.

 B3 Labs blue brand artwork

B3.fun B3IQ extends B3 with verifiable AI compute supply, node onboarding, and proof surfaces.

 B3 3D isomorphic network mark

B3 Network Explorer Broader B3 network context lives on the explorer while B3IQ keeps route readiness here.

What Exists Now

Local node stack

Cross-platform host agent, setup wizard, local owner auth, local API keys, model profile readiness, and local OpenAI-compatible chat.

Learn More
Hosted gateway foundations

Public /v1/api/models, /v1/api/stats, /v1/api/pricing, authenticated /v1/api/chat/completions, customer-owned gateway keys, own-nodes-strict routing policy, streaming settlement metadata, and redacted accounting.

Learn More
Receipts and payouts

Signed receipts, per-key/account usage metering, node payout rows, payout batches, and public-safe settlement views.

Learn More
Protocol explorer

Public views for nodes, models, route readiness, receipts, settlement, and protocol classes.

What Must Stay Out Of Scope

Endpoint-first discipline

B3IQ should not maintain product-specific client libraries, agent SDKs, or bespoke adapters. The supported developer surface is endpoint behavior, OpenAPI contracts, examples, error codes, and public protocol metadata.

Launch Gate For The Vision

The vision is credible only when B3IQ proves real routed work without hiding gaps behind roadmap language:

Vision launch gate
Protocol-ready pilot supply

Protocol-eligible, benchmarked, route-ready GPU supply exists in the current pilot path.

Real routed work evidence

Hosted completions, generations, streaming, and private process work exercise real routed traffic and produce recorded evidence.

Accounting evidence

Receipts, usage rows, node payout rows, and public telemetry line up in redacted evidence packets.

Production operating gates

Production installer trust, command-center controls, customer auth, and incident controls are complete enough for public traffic.

Read More

Roadmap

Current capabilities, open launch gates, and production sequencing.

Learn More
Protocol overview

How proof, settlement, and reputation fit together.

Learn More
Private inference

The active privacy boundary and future confidential compute path.

Learn More
Previous Local, Managed, Network
Next Roadmap
Ask a question... ⌘I