Models 4/10OneAI configuredopenai · gpt-4o-miniAI expandingP6 watch

A living memory and coordination layerfor networked AI.

OneField turns live signals into shared memory, routes context to any model, and returns evidence-backed coordination for humans, agents, and networks.

models

4/10

ready providers

builders

143

recognized nodes

relations

75

coordination edges

signals

8

AI-readable inputs

143 builders75 relationsavg 69
WAOC

Builder

Cryptoliver

Builder

Lee

Builder

MGR ᴄʜᴏᴘᴘᴀ

Builder

STEVO

Builder

Hamz ||

Builder

说的 道理

Builder

Mer

Builder

MissHana🌿

Builder

signal pulse active
Model mesh4/10

providers ready

Default modelgpt-4o-mini

routing target

OneAIconfigured
Live loopready
Proposals9
Signal themes3
P6watch

AI flow

Signal becomes memory. Memory becomes task.

1

Signal Intake

Market, social, Telegram, OneMission, agent logs.

2

Memory Graph

Signals become evidence-backed graph context.

3

Model Mesh

OneAI, OpenAI, OpenRouter, DeepSeek, and future models reason over scoped context.

4

Task Draft

Models return proposals, memory candidates, and task drafts.

5

Trust Compression

Risky actions pass through approval boundaries.

6

Human / Agent Action

Execution happens only after governance.

System Judgment 1-7

watch

From system visibility to system judgment.

System Judgment turns raw signals, memory candidates, task pressure, and risk boundaries into a coordinated AI decision layer.

OneAI decision confidence

48

Judgment API

Signal triage

Agent memory candidate loop ready

Route: background-context

low

Agent memory layer narrative is rising

Route: background-context

low

BTC market volume expansion

Route: background-context

low

Memory lifecycle

Task recommendations

Mission backlog pressure

Owner: Field operator

high

Network density

Owner: Field operator

medium

Trust baseline

Owner: Field operator

medium

Risk judgment

Data privacy risk

Use scoped context policies and redaction before model calls.

watch

Financial / trading risk

Keep OKX and Binance public-readonly until a separate asset governance layer exists.

watch

Social coordination risk

Require evidence and owner confirmation before sensitive introductions.

watch

OneAI blocked actions

direct memory writemission creation without approvaltrading or withdrawalprivate data exportrole escalation

Network coordination feed

Agent memory candidate loop ready

agent signal routed to background-context.

new-signal

Agent memory layer narrative is rising

x signal routed to background-context.

new-signal

BTC market volume expansion

binance signal routed to background-context.

new-signal

BTC-USDT swap open interest

okx signal routed to background-context.

new-signal

P8 Production Bridge

blocked

The production gaps are now explicit launch lanes.

P8 Production Bridge turns OneField's missing production gaps into explicit launch lanes with gates, endpoints, no-data-loss boundaries, and next actions.

Bridge score

19

Bridge API

Production Write Path

watch

WriteIntent creation exists but remains gated by database mode, backup, and ONEFIELD_P6_WRITES_ENABLED.

Use /api/production/write-intents/dry-run before any real POST, then create only targeted intents.

User Register / Claim / Permission Closure

watch

Session signing is production-ready, but durable claim and role request tables still need an additive migration.

Add additive UserClaim, RoleUpgradeRequest, and Invite tables before enabling register/claim writes.

OneAI Live Task / Proposal Loop

ready

OneAI live proposal infrastructure is ready or reachable.

Use /api/oneai/health?probe=true, then POST /api/oneai/live-proposals from an admin/steward session.

OKX / Binance / X / Telegram Realtime Signal Flow

ready

4 signal sources are configured/live; 0 sources report error.

Keep exchanges public-readonly; add scheduler/webhook replay checks before treating feeds as realtime production.

Versioned Long-term User Memory

blocked

FieldMemoryVersion exists, but full memory create/revoke/export lifecycle still needs dedicated routes and policy checks.

Enable memory writes target-by-target after backup; start with versioned memory create and export APIs.

Observability / Rate Limit / Billing / Tenant Boundary

blocked

Basic in-memory rate limiting and audit logging exist; tenant, billing, usage ledger, and external observability are not complete.

Add additive UsageEvent, Tenant, ApiKey, BillingAccount, and SecurityReview tables before opening external customers.

Security Review and Launch Gate

blocked

Security launch review requires admin/steward context.

Run the launch checklist and rotate any exposed keys before opening real external access.

p6WritesEnabled: falseoneaiLiveEnabled: trueoneaiKeyConfigured: trueauthSecretReady: true

Model Mesh

All-model memory router

4/10

OneField can route scoped memory and coordination context to OneAI, OpenAI, OpenRouter, DeepSeek, Claude, Gemini, Grok, Qwen, local, or custom models.

Open layer

Workbench

Context to drafts

preview

Models can process OKX/Binance/X/Telegram signals, Memory Graph, missions, builders, risks, and coordination opportunities into safe drafts.

Open layer

Review

Draft promotion gate

3

AI drafts now pass evidence, confidence, risk, scope, and write-boundary gates before becoming reviewable memory or coordination candidates.

Open layer

Intent

Write preview

1

Promotable drafts are shaped into non-submitting P6 write-intent previews with backup, approval, and rollback requirements.

Open layer

OneAI

Live proposal loop

ready

Real OneAI calls can be enabled on demand, then every output is validated before becoming a draft.

Open layer

Signals

Fusion engine

3

Market, social, mission, and agent signals are fused into semantic themes for AI reasoning.

Open layer

Memory

Memory Graph v2

40

Builders, evidence, missions, risks, and proposals are connected into a queryable memory graph.

Open layer

P6

Controlled writes

watch

AI can propose freely, but real writes require intent, approval, audit, backup, and rollback.

Open layer

Architecture Map

How OneField connects models, signals, memory, and tasks

The live graph now leads the homepage. This architecture panel shows the larger memory OS shape around it, including model routing.

strongest edge

95

relationship strength

visible nodes

9

homepage topology

visible edges

13

active graph preview

OneField
OneAI
Signals
Memory
Tasks
OneAI configured9 proposals

Memory OS routing

143 builders, 8 fused signals, and 9 AI proposals are kept as separated but connected layers.

Current AI task surface

From model reasoning to proposal

Review Agent Gateway

agent node has score 35, 10 graph links, and 1 evidence references.

high

Model output normalizer is stable

8 fields define the shared output boundary for model-generated insights, memory candidates, and task drafts.

normalizer

Write-intent drafts are non-committing

6 drafts can be inspected, but none submit automatically.

no commit

Operating model

Expand first. Compress at the boundary.

Generate memory proposals

Allowed in expansion mode while it remains proposal-only.

expand

Generate mission drafts

Allowed in expansion mode while it remains proposal-only.

expand

Explore builder connections

Allowed in expansion mode while it remains proposal-only.

expand

Suggest circle activation

Allowed in expansion mode while it remains proposal-only.

expand

Durable memory writes

Must pass Trust Compression before any real action.

compress

Mission creation or external dispatch

Must pass Trust Compression before any real action.

compress

Graph relation mutation

Must pass Trust Compression before any real action.

compress

Profile or identity change

Must pass Trust Compression before any real action.

compress

Signal fusion

What models can read now

Market narrative to memory

Market signals can shape ecosystem memory and timing, but not trading instructions.

4 signals

Social narrative to coordination

Public narrative can reveal demand, confusion, momentum, or builder alignment opportunities.

2 signals

Agent reflection to proposal

Agent and mission signals can become proposals for memory, missions, and operating cadence.

2 signals

OneAI health

OneAI is configured. Run ?probe=true from production to verify live reachability.

Base URL

https://oneai-api-production.up.railway.app is configured.

pass

API key

ONEAI_API_KEY is configured. Value is hidden.

pass

Live proposal flag

POST /api/oneai/live-proposals may call OneAI when authenticated.

pass

Network

Add ?probe=true to test network and OneAI generate reachability.

skipped
Run probe API

WAOC core infrastructure

OneField is becoming the living memory and coordination layer between humans, agents, models, and networks.

The homepage now reflects the real architecture: Model Mesh, OneAI live preview, AI expansion, signal fusion, governed memory, and P6 controlled writes. This is the front door to the system, not just a brochure.