diff --git a/docs/architecture-overview.svg b/docs/architecture-overview.svg index 94ff1b1..fbc3a4e 100644 --- a/docs/architecture-overview.svg +++ b/docs/architecture-overview.svg @@ -1,87 +1,101 @@ - - Nexus One AI Architecture Overview - Business-style architecture figure showing entry point, experience layer, core services, control surfaces, and deployment substrate for Nexus One AI. + +Nexus One AI Architecture Overview +Business-style architecture figure showing entry point, experience layer, core services, control surfaces, and deployment substrate for Nexus One AI, scaled across commercial tiers. + - - + + + + + + + + + - + + - NEXUS ONE AI · ARCHITECTURE OVERVIEW - Private AI Platform on Customer-Controlled Infrastructure - Portal layer, core services, and control surfaces packaged for on-premises deployment. - + + + NEXUS ONE AI · ARCHITECTURE OVERVIEW + Private AI Platform on Customer-Controlled Infrastructure + Portal layer, core services, and control surfaces packaged for on-premises deployment. + - ENTRY POINT - EXPERIENCE LAYER - CORE SERVICES - CONTROL SURFACES - - - Users - Admins - Department teams - Procurement / IT / Security - - - - Portal and Access Layer - Nexus One AI Portal - Open WebUI - Authentication and session control - Prompt workflows and admin UI - - - Inference - Ollama - Local model runtime - - - Knowledge Retrieval - ChromaDB - Upload and RAG pipeline - - - Control Plane - FastAPI backend - Audit, backup, settings - - - Observability - Grafana and Prometheus - GPU and system metrics - - DEPLOYMENT SUBSTRATE - - Ubuntu 22.04 • Bootable ISO or scripted installer • CPU or NVIDIA GPU infrastructure • Internal network / air-gapped friendly - - - - - - - - - - - - Private AI packaged stack - Powered by Cezen - + +SCALES BY COMMERCIAL TIER +Starter +1–5 users · NVIDIA RTX 5090, 32GB VRAM +Basic +5–20 users · NVIDIA RTX Pro 6000, 96GB VRAM + +Pro +20–100 users · RTX Pro 6000 Blackwell, 288GB + +Max +100+ users · NVIDIA HGX H200, 1.1TB VRAM + + +ENTRY POINT + + +Users +Admins, department +teams, and +procurement / IT / +security reviewers. + +EXPERIENCE LAYER + + +Portal and Access Layer +Nexus One AI Portal and Open WebUI +Authentication and session control. +Prompt workflows and admin UI. + +CORE SERVICES + + +Inference +Ollama, plus vLLM on Pro/Max +Local model runtime — CPU or NVIDIA +GPU. + + +Knowledge Retrieval +ChromaDB +Document upload and RAG pipeline. + +CONTROL SURFACES + + +Control Plane +FastAPI backend — 156 routes +Audit, backup, settings, and +license/tier enforcement. + + +Observability +Grafana and Prometheus +GPU and system metrics. + + + + + +DEPLOYMENT SUBSTRATE + +Ubuntu 22.04 · bootable ISO or scripted installer +CPU or NVIDIA GPU infrastructure · internal network / air-gapped friendly + + +Private AI packaged stackPowered by Cezen + \ No newline at end of file diff --git a/docs/nexus-one-ai-opportunity.jpg b/docs/nexus-one-ai-opportunity.jpg index 0948324..3635b5a 100644 Binary files a/docs/nexus-one-ai-opportunity.jpg and b/docs/nexus-one-ai-opportunity.jpg differ diff --git a/docs/nexus-one-ai-package-overview.jpg b/docs/nexus-one-ai-package-overview.jpg index 31b5ca3..3091ab3 100644 Binary files a/docs/nexus-one-ai-package-overview.jpg and b/docs/nexus-one-ai-package-overview.jpg differ diff --git a/docs/nexus-one-ai-tier-comparison.jpg b/docs/nexus-one-ai-tier-comparison.jpg index 22bb29a..65d1176 100644 Binary files a/docs/nexus-one-ai-tier-comparison.jpg and b/docs/nexus-one-ai-tier-comparison.jpg differ diff --git a/docs/technical-service-flow.svg b/docs/technical-service-flow.svg index b46084c..015fced 100644 --- a/docs/technical-service-flow.svg +++ b/docs/technical-service-flow.svg @@ -1,140 +1,175 @@ - - Nexus One AI Technical Service Flow - Technical service flow diagram for Nexus One AI showing user access, portal services, backend orchestration, inference, retrieval, storage, and operations components inside customer-controlled infrastructure. + +Nexus One AI Technical Service Flow +Technical service flow diagram for Nexus One AI showing user access, portal services, backend orchestration, inference, retrieval, storage, and operations components inside customer-controlled infrastructure, scaled across commercial tiers. + - - + + + + + + + + + - - + + - NEXUS ONE AI · TECHNICAL SERVICE FLOW - Runtime, Data, and Operations Paths Inside Customer-Controlled Infrastructure - Separates user access, orchestration, retrieval, inference, storage, and support services for buyer evaluation and technical review. - + + + NEXUS ONE AI · TECHNICAL SERVICE FLOW + Runtime, Data, and Operations Paths Inside Customer-Controlled Infrastructure + Separates user access, orchestration, retrieval, inference, storage, and support services for buyer evaluation and technical review. + - - ACCESS SURFACES - - - APPLICATION AND ORCHESTRATION - - - AI AND DATA SERVICES - - - OPERATIONS AND PLATFORM - - - Users - Admins and review teams - Department users - Security and procurement - - - Portal Access - Browser over internal network - Nexus One AI portal - Open WebUI surface - - - Integration Access - API consumers - Service accounts - Internal automation jobs - - - Portal and Session Layer - Nexus One AI Portal and Open WebUI - Authentication, session control, and admin UX - Prompt workflows, model selection, and review screens - - - FastAPI Control Plane - REST APIs and orchestration - License, settings, users, API keys, and system reports - Audit logs, backup and restore, readiness, and install state - Feature gating by commercial tier - - - Workflow and Agent Execution - Server-side task coordination - Prompt chains, scheduled jobs, document actions - Connector sync and save-to-knowledge-base flows - - - Inference Runtime - Ollama local model serving - Model catalog, runtime status, and local inference - CPU or NVIDIA GPU execution path - - - Knowledge Retrieval - ChromaDB and embedding pipeline - Document upload, chunking, indexing, and query retrieval - RAG context delivery back to application services - - - Content and Config Storage - Persistent local state - Uploads, backups, install config, and license artifacts - Model files and service configuration - - - Observability - Prometheus and Grafana - System, GPU, and service health metrics - Operational dashboards for support and handover - - - Deployment Substrate - Ubuntu 22.04 install target - Bootable ISO or scripted installer path - Internal-network and air-gapped deployment friendly - - - Operations Controls - Backup, restore, readiness, and audit - Pre-restore safety snapshots and support workflows - Commercial licensing and post-install administration - - - - - - - - - - - - - - - - - - Customer-controlled boundary: portal, APIs, inference, retrieval, and operational telemetry remain on infrastructure owned or managed by the customer. - Powered by Cezen - + +SCALES BY COMMERCIAL TIER +Starter +1–5 users · NVIDIA RTX 5090, 32GB VRAM +Basic +5–20 users · NVIDIA RTX Pro 6000, 96GB VRAM + +Pro +20–100 users · RTX Pro 6000 Blackwell, 288GB + +Max +100+ users · NVIDIA HGX H200, 1.1TB VRAM + + +ACCESS SURFACES + + +Users +Admins and review teams +Department users, +security, and +procurement +reviewers. + + +Portal Access +Browser over internal network +Nexus One AI portal +and the Open WebUI +chat surface. + + +Integration Access +API consumers +Service accounts and +internal automation +jobs. + +APPLICATION AND ORCHESTRATION + + +Portal and Session Layer +Nexus One AI Portal and Open WebUI +Authentication, session control, and +admin UX. +Prompt workflows, model selection, +and review screens. + + +FastAPI Control Plane +REST APIs and orchestration — 156 routes +License, tier, users, API keys, and +system reports. +Model router and guardrails on +Pro/Max tiers. +Audit logs, backup/restore, and +readiness state. + + +Workflow and Agent Execution +Server-side task coordination +Agent builder, scheduled jobs, and +document actions. +Connector sync and save-to-knowledge- +base flows. + +AI AND DATA SERVICES + + +Inference Runtime +Ollama local model serving +Model catalog, runtime status, and +local inference. +vLLM OpenAI-compatible serving on +Pro/Max tiers. +CPU or NVIDIA GPU execution path. + + +Knowledge Retrieval +ChromaDB and embedding pipeline +Document upload, chunking, +indexing, and query retrieval. +RAG context delivery back to +application services. + + +Applied AI Services +Document and meeting intelligence +Document parsing, summarization, +and meeting transcription. +Evaluation suite and QLoRA fine- +tuning on Pro/Max. + + +Content and Config Storage +Persistent local state +Uploads, backups, install config, +and license artifacts. +Model files and service +configuration. + +OPERATIONS AND PLATFORM + + +Observability +Prometheus and Grafana +System, GPU (DCGM), and service +health metrics. +Operational dashboards for support +and handover. + + +Deployment Substrate +Ubuntu 22.04 install target +Bootable ISO or scripted installer +path. +vLLM, JupyterLab, MLflow, and +MinIO on higher tiers. +Internal-network and air-gapped +deployment friendly. + + +Operations Controls +Backup, restore, readiness, and audit +Feasibility and readiness reports +for handover. +Commercial licensing and tiered +feature gating. + + + + + + + + + + + + + +Customer-controlled boundary: portal, APIs, inference, retrieval, and operational telemetry remain on infrastructure owned or managed by the customer.Powered by Cezen + \ No newline at end of file