1---
2name: deploy-model
3description: "Unified Azure OpenAI model deployment skill with intelligent intent-based routing. Handles quick preset deployments, fully customized deployments (version/SKU/capacity/RAI policy), and capacity discovery across regions and projects. USE FOR: deploy model, deploy gpt, create deployment, model deployment, deploy openai model, set up model, provision model, find capacity, check model availability, where can I deploy, best region for model, capacity analysis. DO NOT USE FOR: listing existing deployments (use foundry_models_deployments_list MCP tool), deleting deployments, agent creation (use agent/create), project creation (use project/create)."
4license: MIT
5metadata:
6  author: Microsoft
7  version: "1.0.0"
8---
9 
10# Deploy Model
11 
12> **Scope — read this first.** This skill creates model deployments **out-of-band** via Azure CLI / MCP / portal. For azd-managed Foundry projects (those scaffolded from `azd-ai-starter-basic` or via `azd ai agent init`), declare deployments in `azure.yaml services.<name>.config.deployments[]` instead — `azd ai agent init` writes the entry from the sample manifest and `azd provision` creates the deployment through Bicep. See [foundry-agent/create/create-hosted.md](../../foundry-agent/create/create-hosted.md) for the Golden Path. Use this skill only for: (a) Foundry projects not managed by an azd project, (b) ad-hoc deployments outside the azd lifecycle.
13 
14Unified entry point for all Azure OpenAI model deployment workflows. Analyzes user intent and routes to the appropriate deployment mode.
15 
16## Quick Reference
17 
18| Mode | When to Use | Sub-Skill |
19|------|-------------|-----------|
20| **Preset** | Quick deployment, no customization needed | [preset/SKILL.md](preset/SKILL.md) |
21| **Customize** | Full control: version, SKU, capacity, RAI policy | [customize/SKILL.md](customize/SKILL.md) |
22| **Capacity Discovery** | Find where you can deploy with specific capacity | [capacity/SKILL.md](capacity/SKILL.md) |
23 
24## Intent Detection
25 
26Analyze the user's prompt and route to the correct mode:
27 
28```
29User Prompt
30    │
31    ├─ Simple deployment (no modifiers)
32    │  "deploy gpt-4o", "set up a model"
33    │  └─> PRESET mode
34    │
35    ├─ Customization keywords present
36    │  "custom settings", "choose version", "select SKU",
37    │  "set capacity to X", "configure content filter",
38    │  "PTU deployment", "with specific quota"
39    │  └─> CUSTOMIZE mode
40    │
41    ├─ Capacity/availability query
42    │  "find where I can deploy", "check capacity",
43    │  "which region has X capacity", "best region for 10K TPM",
44    │  "where is this model available"
45    │  └─> CAPACITY DISCOVERY mode
46    │
47    └─ Ambiguous (has capacity target + deploy intent)
48       "deploy gpt-4o with 10K capacity to best region"
49       └─> CAPACITY DISCOVERY first → then PRESET or CUSTOMIZE
50```
51 
52### Routing Rules
53 
54| Signal in Prompt | Route To | Reason |
55|------------------|----------|--------|
56| Just model name, no options | **Preset** | User wants quick deployment |
57| "custom", "configure", "choose", "select" | **Customize** | User wants control |
58| "find", "check", "where", "which region", "available" | **Capacity** | User wants discovery |
59| Specific capacity number + "best region" | **Capacity → Preset** | Discover then deploy quickly |
60| Specific capacity number + "custom" keywords | **Capacity → Customize** | Discover then deploy with options |
61| "PTU", "provisioned throughput" | **Customize** | PTU requires SKU selection |
62| "optimal region", "best region" (no capacity target) | **Preset** | Region optimization is preset's specialty |
63 
64### Multi-Mode Chaining
65 
66Some prompts require two modes in sequence:
67 
68**Pattern: Capacity → Deploy**
69When a user specifies a capacity requirement AND wants deployment:
701. Run **Capacity Discovery** to find regions/projects with sufficient quota
712. Present findings to user
723. Ask: "Would you like to deploy with **quick defaults** or **customize settings**?"
734. Route to **Preset** or **Customize** based on answer
74 
75> 💡 **Tip:** If unsure which mode the user wants, default to **Preset** (quick deployment). Users who want customization will typically use explicit keywords like "custom", "configure", or "with specific settings".
76 
77## Project Selection (All Modes)
78 
79Before any deployment, resolve which project to deploy to. This applies to **all** modes (preset, customize, and after capacity discovery).
80 
81### Resolution Order
82 
831. **Check `PROJECT_RESOURCE_ID` env var** — if set, use it as the default
842. **Check user prompt** — if user named a specific project or region, use that
853. **If neither** — query the user's projects and suggest the current one
86 
87### Confirmation Step (Required)
88 
89**Always confirm the target before deploying.** Show the user what will be used and give them a chance to change it:
90 
91```
92Deploying to:
93  Project:  <project-name>
94  Region:   <region>
95  Resource: <resource-group>
96 
97Is this correct? Or choose a different project:
98  1. ✅ Yes, deploy here (default)
99  2. 📋 Show me other projects in this region
100  3. 🌍 Choose a different region
101```
102 
103If user picks option 2, show top 5 projects in that region:
104 
105```
106Projects in <region>:
107  1. project-alpha (rg-alpha)
108  2. project-beta (rg-beta)
109  3. project-gamma (rg-gamma)
110  ...
111```
112 
113> ⚠️ **Never deploy without showing the user which project will be used.** This prevents accidental deployments to the wrong resource.
114 
115## Pre-Deployment Validation (All Modes)
116 
117Before presenting any deployment options (SKU, capacity), always validate both of these:
118 
1191. **Model supports the SKU** — query the model catalog to confirm the selected model+version supports the target SKU:
120   ```bash
121   az cognitiveservices model list --location <region> --subscription <sub-id> -o json
122   ```
123   Filter for the model, extract `.model.skus[].name` to get supported SKUs.
124 
1252. **Subscription has available quota** — check that the user's subscription has unallocated quota for the SKU+model combination:
126   ```bash
127   az cognitiveservices usage list --location <region> --subscription <sub-id> -o json
128   ```
129   Match by usage name pattern `OpenAI.<SKU>.<model-name>` (e.g., `OpenAI.GlobalStandard.gpt-4o`). Compute `available = limit - currentValue`.
130 
131> ⚠️ **Warning:** Only present options that pass both checks. Do NOT show hardcoded SKU lists — always query dynamically. SKUs with 0 available quota should be shown as ❌ informational items, not selectable options.
132 
133> 💡 **Quota management:** For quota increase requests, usage monitoring, and troubleshooting quota errors, defer to the [quota skill](../../quota/quota.md) instead of duplicating that guidance inline.
134 
135## Prerequisites
136 
137All deployment modes require:
138- Azure CLI installed and authenticated (`az login`)
139- Active Azure subscription with deployment permissions
140- Azure AI Foundry project resource ID (or agent will help discover it via `PROJECT_RESOURCE_ID` env var)
141 
142## Sub-Skills
143 
144- **[preset/SKILL.md](preset/SKILL.md)** — Quick deployment to optimal region with sensible defaults
145- **[customize/SKILL.md](customize/SKILL.md)** — Interactive guided flow with full configuration control
146- **[capacity/SKILL.md](capacity/SKILL.md)** — Discover available capacity across regions and projects
147