The platform routes tasks intelligently:
● Routine or structured tasks run on a local on-prem LLM.
● Higher-quality foundation models are used only when necessary.
● Persistent vector memory ensures only relevant context is retrieved, reducing unnecessary token load.
You use your own LLM subscriptions, so you only pay for what you consume — and the system minimizes unnecessary external calls.

Most AI tools forget everything between sessions.
 Our platform stores structured memory in a local vector database so that:
● Knowledge persists across runs
● Context improves over time
● Memory entries can be inspected, edited, approved, or deprecated
● Workflows become stable and repeatable
 This turns AI from “chat sessions” into a governed engineering system.

No. The platform is model-agnostic.
You can:
● Use OpenAI, Anthropic, Azure, or other foundation models
● Use a small on-prem LLM for local processing
● Switch providers without rebuilding your workflows
This prevents vendor lock-in and keeps your architecture flexible.

Yes.
 Because data and memory are stored locally, the platform supports:
● Stronger data control
● Reduced cloud exposure risk
● Air-gapped deployment options
● Policy enforcement and audit tracing
● SSO and role-based access control (Enterprise tier)
 This significantly simplifies enterprise security reviews.

No.
 The platform is designed to reduce manual orchestration and repetitive AI setup.
 It includes:
● Reusable workflows
● Built-in evaluation and replay
● Automated routing between models
● Governance and audit logging
 Small teams can deliver production-grade AI systems without building complex infrastructure from scratch.

Requirements depend on the model size you choose.
For small on-prem LLMs used for routine processing, a modern machine with sufficient RAM is typically adequate.
Foundation models still run through your existing cloud subscription when higher-quality output is required.
 You can scale your local model capacity based on your needs.

Chat-based tools are great for experiments.
 This platform is built for engineering systems:
● Persistent memory
● Inspectable vector storage
● Replayable executions
● Governance and policy controls
● Multi-step workflow orchestration
● Enterprise auditability
 It moves AI from experimentation to structured, repeatable, and governed production workflows.

Even though the platform runs on your local machine or internal infrastructure, licensing still applies — because the license governs access to platform capabilities, not data hosting. There are two completely separate licenses involved:

a. License / Subscription for NYEX (The Platform)
This license gives you the right to use the NYEX AI Engineering Platform itself.
What the NYEX license controls:
● Access to the platform software
● Feature tiers (Free / Pro / Team / Enterprise)
● Governance capabilities (policy engine, approvals, audit exports)
● Collaboration features (multi-user workspaces)
● Enterprise controls (SSO, RBAC, compliance modules)
● Certified builds and enterprise distribution (if applicable)
● Software updates and support
 What the NYEX license does NOT include:
● LLM token usage
● Foundation model subscription
● Cloud AI costs
● Local hardware costs
 ● The NYEX license governs the engineering infrastructure layer, not the AI compute layer.

b. License / Subscription for the LLM Provider
This is completely separate.
 You bring your own:
● OpenAI subscription
● Anthropic subscription
● Azure OpenAI subscription
● Or any other foundation model provider
 You pay those providers directly for:
● Token usage
● Model access
● API calls
● Any cloud compute involved

NYEX does not resell tokens and does not add markup to your LLM costs.
This means:
● You pay only for what you consume.
● You maintain full control over your LLM billing.
● You can switch providers without changing your NYEX license.

c. Local LLM Usage
If you use a small on-prem LLM:
You do not pay per token.
Costs are limited to your own hardware.
NYEX simply orchestrates and routes tasks to it.
NYEX provides the orchestration and governance layer — not the model license.

Why This Separation Matters
1. Cost Transparency
You see exactly what you’re paying for:
● Platform capability (NYEX)
● AI compute usage (LLM provider) 

2. No Vendor Lock-in
You can:
● Change model providers
● Adjust usage patterns
● Control token consumption
Without affecting your NYEX license.

3. Enterprise Procurement Simplicity
Enterprises often:
● Already have LLM provider agreements
● Already have cloud contracts
● Already have model governance policies
NYEX integrates into those existing agreements rather than replacing them.