Immersive Reader, Video Indexer and Retired Services

This content was generated with the assistance of AI and should be verified against the official documentation before being relied on in production.

See the section bibliography for the reference material consulted while preparing these pages.

This page targets Azure CLI 2.90.0, the Immersive Reader JavaScript SDK (@microsoft/immersive-reader-sdk), and azure-identity 1.26.0 / Azure.Identity 1.21.0 for the token service, checked on 2026-10-06. Retirement dates come from the official notices linked in the table and were re-checked on 2026-10-06 where the notice was reachable.

Immersive Reader

Azure AI Immersive Reader is an embeddable reading experience. Your web page launches it in a full-screen iframe with the text to read, and it offers:

  • read aloud with word highlighting, in many voices and speeds

  • translation of the whole text or single words into 100+ languages

  • parts of speech coloring (nouns, verbs, adjectives) and syllable splitting

  • a picture dictionary for single words

  • line focus, text spacing and theme options that help readers with dyslexia

Immersive Reader needs its own single-service resource (--kind ImmersiveReader) with a custom subdomain and Entra ID authentication. Your backend gets a token for the scope https://cognitiveservices.azure.com/.default and hands it to the page, together with the subdomain; the page never sees a secret.

Token flow

# catalog-api endpoint: GET /immersive-reader/token
from azure.identity import DefaultAzureCredential

credential = DefaultAzureCredential()

def immersive_reader_token():
    token = credential.get_token("https://cognitiveservices.azure.com/.default")
    return {"token": token.token, "subdomain": "reader-bookshelf"}

Source: Create an Immersive Reader resource and configure Entra authentication

using Azure.Core;
using Azure.Identity;

app.MapGet("/immersive-reader/token", async () =>
{
    AccessToken token = await new DefaultAzureCredential().GetTokenAsync(
        new TokenRequestContext(new[] { "https://cognitiveservices.azure.com/.default" }));
    return Results.Ok(new { token = token.Token, subdomain = "reader-bookshelf" });
});

Source: Immersive Reader quickstart (C#)

The page launches the reader with the SDK:

import { launchAsync } from "@microsoft/immersive-reader-sdk";

const { token, subdomain } = await (await fetch("/immersive-reader/token")).json();
const content = {
  title: "Sample chapter: The Lighthouse Keeper",
  chunks: [{ content: document.querySelector("#sample-chapter").innerText, lang: "en", mimeType: "text/plain" }],
};
await launchAsync(token, subdomain, content, { uiZIndex: 2000 });

Source: Immersive Reader SDK reference

Testing the token endpoint with curl:

curl -s https://catalog.bookshelf.example/immersive-reader/token | jq '{subdomain, token: (.token | .[0:16])}'

Source: Create an Immersive Reader resource

Provisioning

In the Azure portal

  1. Create a resource of type Immersive Reader named reader-bookshelf in rg-bookshelf, with custom subdomain reader-bookshelf.

  2. On its Access control (IAM) blade, assign Cognitive Services Immersive Reader User to id-catalog-api.

With the Azure CLI

az cognitiveservices account create \
  --name reader-bookshelf --resource-group "$RG" --location "$LOCATION" \
  --kind ImmersiveReader --sku S0 --custom-domain reader-bookshelf --yes

az role assignment create \
  --assignee "$(az identity show --name id-catalog-api --resource-group "$RG" --query principalId -o tsv)" \
  --role "Cognitive Services Immersive Reader User" \
  --scope "$(az cognitiveservices account show --name reader-bookshelf --resource-group "$RG" --query id -o tsv)"

Source: Create an Immersive Reader resource

Use cases

Scenario Feature Why this feature Example input → output

Bookshelf: sample chapters

Read aloud, translation, line focus

Readers try a chapter in their own way before buying.

Sample chapter → reader with highlighting and Spanish translation

Education

Parts of speech, syllables, picture dictionary

Early readers and language learners.

Lesson text → colored parts of speech

Accessibility

Read aloud, text spacing, themes

Low-vision and dyslexic readers.

Article → high-contrast spaced view with audio

Dyslexia support

Line focus, syllables

Reduces visual crowding.

Homework text → one line at a time

When not to use it

Azure AI Video Indexer

Azure AI Video Indexer analyzes video and audio (transcripts, faces, labels, topics, scenes, brands) with its own portal, widgets and API. It is a separate service built on Azure AI capabilities, not part of the Foundry Tools catalog.

  • Choose Video Indexer for a media library with a ready-made portal, embeddable player widgets and media-specific insights.

  • Choose Content Understanding video analyzers (Content Understanding: Multimodal Analyzers) when you want your own fields, chapters and summaries as JSON inside a pipeline, next to document and audio analyzers.

Source: What is Azure AI Video Indexer?

Retired and deprecated services

Foundry Tools retirement timeline 2025 to 2029: retired features on the left, deprecated features with end dates on the right

Retired features are listed only here and in short migration notes on the tool pages. Deprecated-but-supported features carry a warning on their own page.

Item Status Date Replacement Notice

QnA Maker

Retired

2025-03-31

Custom question answering

Azure update

Background removal, Product Recognition, Image Analysis model customization

Retired

2025-03-31

None

What’s new in Vision

Pronunciation content assessment

Retired

July 2025

Azure OpenAI (Foundry Models)

Speech release notes

Speaker Recognition

Retired

2025-09-30 (Speech SDK 1.47)

None; diarization is unaffected

Speech release notes

Speech intent recognition

Retired

October 2025 (Speech SDK 1.47)

CLU or an LLM

Speech release notes

Spatial Analysis, Video Retrieval

Retired

2025

Content Understanding (+ AI Search)

What’s new in Vision

ConversationTranslator (multi-device)

Removed

Speech SDK 1.51

Speech translation, Live Interpreter

Speech release notes

LUIS

Retired

2026-03-31

CLU

Migrate from LUIS

Speech to text REST v3.0 and v3.2 previews

Retired

2026-03-31

REST 2025-10-15

Speech to text REST API

Anomaly Detector

Retired

2026-10-01

Fabric anomaly detection

Azure update

Personalizer

Retired

2026-10-01

microsoft/learning-loop

Azure update

Metrics Advisor

Retired

2026-10-01

None

Azure update

Translator v3 BreakSentence, Detect, Dictionary

Not in 2026-06-06 (still in v3.0)

 — 

Language detection; adaptive custom translation

Migration guide

Content Moderator

Deprecated

Retires 2027-03-15

Content Safety

Content Moderator overview

Language Studio

Deprecated

Retires 2027-03-20

Foundry portal

Migration guide

Foundry classic agents

Deprecated

Retire 2027-03-31

New Foundry agents

Agent Service overview

Document Intelligence v2.1

End of support

2027-09-15

v4.0 2024-11-30

What’s new

Entity linking

Deprecated

End of support 2028-09-01

Language NER or Foundry Models

What’s new in Language

Image Analysis 3.2 / 4.0, multimodal embeddings

Deprecated

Retire 2028-09-25

Document Intelligence Read, Face, Content Understanding, Foundry Models

Migration options

Custom Vision

Deprecated

Retires 2028-09-25

AutoML, Foundry Models, Content Understanding classifier

Migration options

Document Intelligence v3.0

End of support

2029-03-30

v4.0 2024-11-30

What’s new

Language legacy features (sentiment, key phrases, summarization, CLU, CQA, custom text classification, orchestration)

Deprecated

End of support 2029-03-31

Foundry Models

What’s new in Language

The Personalizer date follows the official Azure update (2026-10-01). Some Learn pages showed an earlier date (2026-08-25); either way the service is retired as of this page’s baseline date.

Clean up

az cognitiveservices account delete --name reader-bookshelf --resource-group "$RG"

Source: Azure CLI reference — az cognitiveservices account delete

Related pages

References