Platform | EZ4YouTech.com

How the platform works

Three roles: one isolated company workspace per business. Everyday common utility tools plus one industry pack; depth grows by plan tier. Browse apps by industry.

Platform operator

Creates client workspaces from the operations console
Sets industry vertical and plan (Starter / Standard / Elite / Enterprise)
Controls which catalog apps are Live vs coming soon
Provisions company admins and team sign-ins

Client admin (setup)

One-time AI account setup: connect supported providers under your AI provider account (Bring Your Own Subscription: your API key, your provider bill), including OpenAI, Together, Groq, xAI (Grok), and others; full list in admin setup
Encrypted credentials stored per company workspace; never logged
No day-to-day workspace or industry apps on this login

Team members (daily use)

Run structured AI apps: forms and tuned prompts, not open chat
Plan tier controls user count and which apps unlock
Upload, draft, summarize, and extract inside company workspace boundaries

Security at a glance

Built for client-facing and regulated work. We describe an enterprise-ready architecture. We do not claim SOC 2, ISO 27001, or HIPAA unless agreed in a signed contract.

your AI provider account (Bring Your Own Subscription): your API keys, encrypted per company; never logged; billed by your provider
Tenant isolation: tenant_id on every JWT-backed request and storage path
Role separation: admins configure keys; users run apps without handling secrets
No data resale: subscription revenue only; prompts not sold for ads or model training

Core security controls

How we protect tenant data and credentials in the live application.

Authentication & access

JWT with tenant_id, plan, and role on every API call
bcrypt password hashing; login rate limiting per IP
RBAC: platform operator, partner, company admin, user
Plan gating: app catalog and agent seats enforced server-side

provider account & data handling

Fernet-encrypted provider keys in MongoDB Atlas
Keys decrypted only for outbound provider requests
Uploads and run history under tenant_id/ storage paths
Production security headers (HSTS, X-Frame-Options, Referrer-Policy)

Deeper diagrams: Architecture & security guide · Security FAQ

Technology stack & operations

For IT and security reviewers: USA production posture (May 2026).

Azure production stack: Container Apps + Key Vault + Blob Storage
MongoDB Atlas: tenants, users, usage, and encrypted subscription credentials
provider account routing: tenant-owned keys to supported providers (try-order when multiple are configured)
Tenant isolation: enforced by tenant_id across API and storage paths

Production stack

Layer	Technology	Role
Edge	HTTPS, CDN	TLS for marketing site and platform hostname
Application	FastAPI + Next.js	REST API and agent/admin workspace UI
Compute	Azure Container Apps	Managed runtime, tagged releases
Secrets	Azure Key Vault	JWT key, encryption key, Mongo URI, not in git
Database	MongoDB Atlas (M10)	Tenants, users, encrypted subscription credentials, usage
Storage	Azure Blob	Tenant-scoped documents and artifacts
AI	OpenAI, Together, Groq, xAI (Grok), DeepSeek, Mistral, Fireworks, OpenRouter, Azure OpenAI	Customer keys; billed on your provider account

Security & data protection

Area	Practice
Data in transit	TLS 1.2+ end-to-end
Secrets at rest	Fernet encryption for AI subscription credentials
Tenant isolation	API checks + `tenant_id/` blob paths
Logging	No keys, passwords, or raw prompts in app logs
Ops	Azure Monitor; Atlas backups; blob soft-delete

Your responsibilities

Protect admin credentials and rotate AI subscription credentials per your policy
Set training / opt-out in your AI provider console
Human-review client-facing outputs in regulated industries

Supported AI providers

Connect the AI services you already use. You pay providers directly; EZ4YouTech.com does not markup tokens.

Your company admin configures credentials and try-order in the Secure AI Platform setup tab (comparison table with illustrative list prices). Router-ready providers:

OpenAI, Together AI, Groq, Fireworks AI, DeepSeek, Mistral AI
xAI (Grok): different vendor from Groq (groq.com)
OpenRouter, Azure OpenAI

Anthropic, Google Gemini, and AWS Bedrock appear in the setup comparison; native API routing ships in a later release. Legacy Anyscale keys are not offered for new setups.

FAQ: supported providers · Open live platform setup