skills.rest
sohamda
sohamdaCOMMUNITY
Explore all skills by this author →
Website𝕏@iamsohamGitHub

customize

Deploys Azure OpenAI models through an interactive guided workflow with full control over version, SKU, capacity, and content filtering.

Updated May 12, 2026
Search Tags:#model deployment#capacity planning#content filtering#azure cli#azure openai#ai foundry#sku selection
One-click install
npx skills add https://github.com/sohamda/apex-try-out-demo-repo --skill customize-sohamda
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: customize
Source: https://github.com/sohamda/apex-try-out-demo-repo/tree/main/.archive/_archived_skills/microsoft-foundry/models/deploy-model/customize
Command: npx skills add https://github.com/sohamda/apex-try-out-demo-repo --skill customize-sohamda

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve? Deploying Azure OpenAI models requires navigating many configuration choices—model versions, SKUs, capacity limits, RAI policies, and advanced options—where mistakes lead to quota errors, failed deployments, or misconfigured production workloads. This Skill walks you through each decision with live quota and capacity data so deployments succeed on the first attempt. ## Core Features & Use Cases - Guided 13-phase deployment flow: Step-by-step selection of model version, SKU (GlobalStandard, Standard, ProvisionedManaged, DataZoneStandard), capacity, and RAI policy, all validated against live Azure quota data. - Cross-region capacity fallback: Automatically queries all Azure regions for available capacity when the current region has none, and helps you switch to a project in a region with quota. - Advanced deployment options: Configure dynamic quota, priority processing for PTU deployments, spillover to backup deployments, and version upgrade policies. - Multi-format model support: Handles OpenAI models via Azure CLI and Anthropic models via ARM REST API with required modelProviderData (industry, country code, organization name). - Use Case: You need to deploy gpt-4o for a production workload with 50K TPM, dynamic quota, and a spillover target. The skill queries your subscription's live quota, validates capacity ranges, and executes the deployment with monitoring until it succeeds. ## Quick Start Ask the AI to deploy an Azure OpenAI model with custom settings, for example: deploy gpt-4o with GlobalStandard SKU and 50K TPM capacity to my Foundry project.

Frequently Asked Questions about customize

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I deploy an Azure OpenAI model with a specific SKU and capacity?▼

Use the guided deployment flow to select your model version, choose from deployable SKUs (GlobalStandard, Standard, ProvisionedManaged), and enter an exact capacity value validated against live quota. The skill queries your subscription's available quota and only presents SKUs you can actually deploy.

What is the difference between GlobalStandard and ProvisionedManaged SKU?▼

GlobalStandard is TPM-based pay-per-use suited for production with high availability, while ProvisionedManaged (PTU) provides reserved guaranteed throughput at fixed monthly cost for predictable high-volume workloads. PTU capacity ranges from 50 to 1000 units in steps of 50.

What happens if my Azure region has no capacity for the model?▼

The skill automatically queries all Azure regions for available capacity of your chosen model, version, and SKU. It presents regions sorted by available capacity and helps you switch to an existing AI Foundry project in a region with quota.

Can I deploy Anthropic models like Claude to Azure AI Foundry?▼

Yes, Anthropic models are deployed via the ARM REST API with required modelProviderData including industry, country code, and organization name fetched from your tenant. Anthropic deployments use capacity of 1 under MaaS billing and do not use RAI policies.

Why does my Azure OpenAI deployment fail with QuotaExceeded?▼

QuotaExceeded occurs when requested capacity exceeds your subscription's available quota for that SKU and model. Check usage with az cognitiveservices usage list, reduce capacity, try a different SKU or region, or request a quota increase.

When should I use customize instead of the preset deployment skill?▼

Use customize when you need precise control over version, SKU, capacity, RAI policy, or advanced options like PTU and spillover. Use preset for quick deployments where the latest version, GlobalStandard SKU, and automatic capacity calculation are acceptable.

Related Skills

Explore More Software Engineering Skills→
microsoftmicrosoft
1.4k

preset

Deploys Azure OpenAI models to optimal regions by analyzing multi-region capacity.

Official
Advanced
dahatakedahatake
2

appinsights-instrumentation

Integrate Azure Application Insights SDKs into ASP.NET Core, Node.js, or Python web apps.

Community
Intermediate
microsoftmicrosoft
1.4k

azure-quotas

Checks Azure quota limits, usage, and available capacity across resource providers and regions.

Official
Intermediate
rrbandarrbanda

preset

Deploy Azure OpenAI models to optimal regions using capacity analysis.

Community
Advanced