Azure OpenAI Services - The Complete Guide

Explore ChatGPT, the leading OpenAI service, and learn what it is, alongside its sibling AI tools like DALL-E, in the introduction to the course.

Explore how ChatGPT functions as an ai chatbot and large language model trained on billions of webpages with human feedback, using prompts and completions in conversational context.

Trace the evolution of OpenAI’s GPT models from the 2015 nonprofit foundation to today’s GPT-5, covering GPT-1 through GPT-4, DALL-E, Codex, and ChatGPT breakthroughs, plus the Microsoft investment.

Understand how tokens encode text as numerical representations that power ChatGPT's predictions and token limits of 4,096 tokens. Explore azure billing by token and tokenizer demonstrations.

Open chatgpt.com, sign up, and configure settings to tailor appearance, language, voice, and data controls. Compare free versus business version, and learn to use chats, sources, and read aloud features.

Master the basic use of ChatGPT for information, translation, and summaries; edit prompts, switch between chats, maintain context, and plan a winter itinerary in Greece.

Learn to work with documents and images in ChatGPT by uploading pdfs like Boeing 747.pdf or photos such as bird.jpeg, then summarize content or describe visuals.

ChatGPT’s image generation uses the Dall E 3 model, demonstrated with an astronaut riding a comet. Learn about timing, free-version limits, and upgrading to the business version for more capacity.

Explore prompt engineering to refine prompts and guide ChatGPT toward precise, high-quality completions. Learn how basic prompts fall short and how targeted prompts reveal exactly what you want.

Learn to craft prompts using the four elements—instruction, context, input data, and output indicator—with the instruction as a mandatory guide for what to do in ChatGPT.

Discover how prompt elements, such as adding context and output indicators, improve results in Azure ChatGPT and OpenAI service through practical formatting and tailored prompts.

Specify prompt engineering to improve ChatGPT results by outlining requirements, using instruction lists and constraints, and requesting details like hotel name, address, price, floors, and proximity to Central Park.

Explore zero-, one-, and few-shot prompting to shape ChatGPT by providing examples. Use a new chat, draft a thunderstorm tweet, and add multiple examples to imitate your voice.

Apply prompt engineering to an exercise, crafting and refining a resignation letter with ChatGPT by adding context and instructions for a concise, respectful result.

Access OpenAI language models hosted in Azure via a REST API, gaining improved reliability, secure deployments with Azure infrastructure and controls, and a 99.9% SLA.

Compare Azure OpenAI service’s 99.9% SLA and security with public options, and explore use cases from customer service chatbots to proprietary data models via Azure OpenAI REST API.

Apply responsible ai practices by following the responsible ai initiative, guidelines, and filters to prevent misuse and blocked responses; ensure OpenAI services uphold 'be nice' behavior.

Understand Azure OpenAI quotas and limits, including per-minute requests: Davinci 120 rpm, ChatGPT 300 rpm, GPT-4 18 rpm, token caps, DALL-E concurrency, and region resource limits.

Learn Azure OpenAI pricing per 1 million tokens, detailing input and output costs for text and image models (GPT-5.1 chat and GPT-Image-1 global) and how token counts affect billing.

Deploy Azure OpenAI Service in the Azure portal by creating a new resource group in East US, naming memi-openai, and selecting the Standard SO pricing tier for OpenAI API access.

Explore how to work with OpenAI text models hosted in Azure, from deployment to testing and using the API. Learn how to complete deployments in the Foundry Portal.

Explore the Microsoft Foundry Portal, a visual tool for deploying and testing AI models, fine-tuning them, and configuring content filters from the Azure portal.

Deploy and test Azure OpenAI models using the Foundry portal, exploring the model catalog, deployments, and the GPT 5.1 Chat option, with global standard deployment and token capacity considerations.

Test deployments in the playground by selecting the active deployment, reviewing system prompts and token limits, and generating a meta prompt for a marketing writing assistant.

Learn to use Azure OpenAI Services API with Postman, including API key authentication and Entra ID options, and build chat prompts while tracking tokens and content filters.

Deploy a chat web app from an Azure OpenAI deployment using a GPT-4.1 model, with content filters and organizational data sources, via the Foundry portal as an app service.

Define content filter levels for Azure OpenAI models using the Azure AI Content Safety Engine, then apply these safeguards in your API to prevent harmful responses.

Explore how to configure content filters in Azure OpenAI Studio, including setting categories like violence, hate, sexual and self-harm with severity levels, and attaching filters to deployments for controlled outputs.

Demonstrates using the Azure OpenAI completion API with content filters in Postman, showing a harmful prompt being filtered and producing a 400 bad request due to content management policy.

Understand quotas and tokens per minute per model per region, and learn how to request a quota increase via the Foundry portal by clicking Request Quota and submitting for review.

Explore using image models in the Azure OpenAI service to generate images from prompts. See how the Foundry portal and the API expose these capabilities.

Deploy DALL-E three in the Foundry portal's images playground, select the correct AI resource and region, then generate and refine beach sunset images from prompts, noting no memory context.

Learn to generate images with the Azure OpenAI image API using Postman, including endpoint discovery, API key setup, JSON body, and customizing size, style, and quality to compare results.

Learn how to use OpenAI on your own data to build chatbots that retrieve information from sources like Azure AI search, blob storage, Cosmos DB, Elasticsearch URL, and file.

Prepare the Azure environment for Azure OpenAI by deploying an AI search resource and storage account, enabling system-assigned identities, and configuring role-based access.

Upload your data in the chat playground, index it with the AI search service, and query the file to reveal that invoices exceeding $1 million require CFO approval.

Learn to use your own data with the API by configuring a data source with Azure Search, endpoint, index, and a system-assigned managed identity, then query for up-to-date results.

Discover how to secure the Azure OpenAI Service using Azure security controls. Apply network security and Azure AD Identity to protect the OpenAI Service.

Secure the OpenAI service network by restricting access to specific networks and blocking public access, using private endpoints and VNets with private IPs.

Deploy a Windows virtual machine in East US 2, connect it to the OpenAI VM VNet, and enable a private endpoint to protect a Foundry resource, disabling public access.

Secure OpenAI access with Entra ID identity instead of an API key, enabling the API to authorize only authorized users and services.

Demonstrate Entra ID authentication to access OpenAI services in Azure Cloud Shell without an API key, by extracting an access token and using a bearer authorization header.

Explore Azure AI Content Safety, a REST API tool that detects hateful, violent, sexual, and self-harm content in real-time to create secure online experiences.

Explore Azure AI Content Safety, which automates text and image moderation, analyzes content in real time with a rest api, and scores for hate speech, violence, sexual content, and self-harm.

Understand pricing for Azure AI content safety: $1 per thousand transactions, 1,024 text characters per transaction, and a 4 MB image limit. Use free tier for learning with limited quotas.

Use the content safety api with postman to test text, configure endpoint, headers, and api key, and explore category analytics for self-harm and violence with severity 0–6.

Learn how Azure AI content safety analyzes images, supporting JPEG, PNG, GIF, BMP, TIFF, WebP up to 4 MB and 50x50–2048x2048, using base64 in Postman or a storage link.

Create a block list named my list and add disallowed words such as knife with a description, then test text analysis by applying the block list to detect matches.

Azure AI content safety studio lets you test text and image moderation, adjust thresholds and block lists, and monitor online activity to fine-tune content safety filters.

Explore how ChatGPT and OpenAI revolutionize software use and simplify planning and summarizing. See how Azure OpenAI Service, Foundry Portal, and security features enable deployment, management, and safe access.