The best LLM in Visual Studio is the one you choose. This month’s update puts that idea front and center: pick the model that fits how you work, then use it everywhere GitHub Copilot already helps you today, from fixing vulnerable packages to reviewing pull requests. And once you’re in the editor, you’ll find sharper debugging tools too, whichever model is doing the thinking alongside you.

Grab the Visual Studio 2026 Stable Channel update and let us know what’s working and where we should invest next.

Updates to Bring Your Own Model

The best AI in Visual Studio is the AI you can bring with you. With Bring Your Own Model (BYOM), more developers can use AI in Visual Studio with the models their teams prefer, whether through GitHub Copilot or their own model deployments. From individual developers experimenting with new models to organizations standardizing on approved deployments, BYOM lets you use AI your way, whether or not you’re signed in to GitHub.

BYOM is available in preview and enabled by default across the Community, Professional, and Enterprise SKUs. Connect to a Microsoft Foundry deployment or another supported provider, including custom endpoints for OpenAI and Ollama, then work in Agent Mode with the new Agent (Preview).

This is an early preview, and we want your feedback. Not every model supports every Agent Mode capability yet. When a feature isn’t supported, Visual Studio marks it clearly instead of failing silently. Try it and tell us what’s working and where we should invest next in the Developer Community forum or fill out this short survey.

To get started: open Chat, open the model picker and choose Add a model or Manage Models, connect to your provider, then start working in Agent Mode. Supported providers are Microsoft Foundry, OpenAI, Anthropic, and Ollama, with custom endpoint support for OpenAI and Ollama.

This release has breaking changes for the BYOM capability: BYOM now works with the new Agent (Preview), which is built on the GitHub Copilot SDK-powered harness. The earlier BYOM experience in the previous Ask and Agent modes is no longer supported; install the latest Visual Studio release and use the new Agent (Preview) to try it. Previously added Ollama models will need to be re-added after this update due to a few backend updates for that provider.

Fix NuGet Vulnerabilities from the Error List

Whichever model you’ve chosen, it can now help you close security gaps the moment they surface. NuGet Audit warnings in the Visual Studio Error List give you a direct way to fix vulnerable packages with GitHub Copilot instead of tracking them down manually.

After a restore reports an NU1901, NU1902, NU1903, or NU1904 warning, select the sparkle action in the Fixer column. This opens GitHub Copilot Chat in agent mode and asks Copilot to fix the vulnerable NuGet packages using the NuGet MCP server.

To use the fix: restore a solution that contains a package with a known vulnerability, open View > Error List and find the NuGet Audit warning, select the sparkle action in the Fixer column, then review and apply the changes Copilot proposes. You’ll need to be signed in to GitHub Copilot, and automated fixes also require the NuGet MCP server, which is built into Visual Studio and can be enabled from the Copilot Chat tools menu.

Explore Pull Requests with the Git Agent

Reviewing a pull request usually means piecing together the discussion thread, the changed files, and what actually matters before you can dig in. The Git agent can now do that legwork for you, powered by whichever model you’ve connected.

Open GitHub Copilot Chat, select the Git agent, and ask it about your pull request, attaching the pull request as context from the Git Repository window. The response includes navigable links that open the corresponding comments and files directly in the pull request view in the IDE, so you can jump straight to what matters instead of scrolling to find it. Ask follow-up questions any time you want the agent to go deeper.

Connect your Git provider’s MCP servers so Copilot can pull in more context about your pull requests. GitHub and Azure DevOps are both supported.

See Which Operand Decides Your if Condition

Ever paused on a compound if statement and had to work out which part of it actually made the call? Visual Studio now highlights the deciding expression in C# if statements that use && or ||.

While debugging, pause on an if statement. Alongside the existing inline → true or → false hint, Visual Studio uses red and green highlighting to show which operand determined the result. When an && condition evaluates to false, the highlighting identifies the operand that failed or caused evaluation to short-circuit. When an || condition evaluates to true, it identifies the operand that satisfied the condition. It’s a small detail that saves you from mentally evaluating each piece of a complex branch on your own.

You can right-click a highlighted expression to disable branch highlighting from the editor context menu, and the highlight colors are available under Tools > Options > Environment > Fonts and Colors if you want to tune them to your theme.

Attach to Process Now Supports Podman Containers

Container discovery now recognizes Podman alongside Docker. Start a container, then go to Debug > Attach to Process. Set Connection type to Podman, then select Find. Your Podman containers appear automatically, with the same attach experience you already use for Docker.

You can debug both .NET and C++ containers with no extra configuration.

From our entire team, thank you for choosing Visual Studio! For the latest updates, resources, and news, check out the Visual Studio Hub and stay in touch.

Happy coding! The Visual Studio team

The post Visual Studio September Update – Power Your Workflow with Your Model appeared first on Visual Studio Blog.

Essential Studio 2026 Volume 3 at a glance

The Essential Studio 2026 Volume 3 expands the Syncfusion toolkit across AI, document processing, forms, UI development, and collaboration. Here are the updates worth exploring.

Release highlights

• JavaScript libraries for Excel and PowerPoint.
• .NET Document Chunking Library for AI and retrieval workflows.
• A2UI for dynamically generated user interfaces.
• WebMCP integration for connecting Syncfusion components with AI assistants.
• Form Builder and Form Renderer for dynamic forms.
• AI-powered PDF experiences.
• New UI controls across .NET MAUI, Pure React, Blazor, and WinUI.
• Collaboration capabilities for documents and spreadsheets.

New AI and AI-ready capabilities

Volume 3 introduces new ways to prepare enterprise content for AI retrieval, generate interfaces dynamically, connect AI assistants with web components, and add intelligent PDF experiences.

Prepare enterprise content for RAG with document chunking

Building a retrieval-augmented generation (RAG) or semantic-search application often starts with preparing documents for retrieval.

The new .NET Document Chunking Library (Preview) transforms PDFs, Word documents, Excel workbooks, PowerPoint presentations, and Markdown files into context-aware chunks while preserving document structure and metadata.

Developers can preserve document hierarchy, control chunk size and overlap, and retain metadata and citations, providing structured content for RAG, semantic search, and knowledge-retrieval workflows.

Create dynamic interfaces with A2UI

AI applications often need more than text responses. They may need forms, charts, grids, or other interactive experiences based on a user’s request.

With A2UI (Preview) for Blazor, React, Angular, Vue, and .NET MAUI, AI agents can dynamically generate interactive user interfaces using Syncfusion components. A2UI components provide the building blocks for these interfaces, while the new A2UI Composer provides a visual way to design and refine them.

Agents describe the interface they need, while the application controls how it is rendered. This enables adaptive experiences such as AI-generated forms, grids, charts, and schedulers without requiring the agent to generate the complete UI implementation.

Connect components with AI Assistants using WebMCP

WebMCP enables AI assistants to interact with Syncfusion React, Angular, JavaScript, JavaScript (ES5), and Vue components through natural language, with component capabilities exposed as tools. Developers can control tool exposure through selection, custom prefixes, and access restrictions.

Build intelligent PDF experiences

The Smart .NET MAUI PDF Viewer adds AI-powered capabilities for applications that work extensively with PDF documents. It can extract document information, detect and populate form fields, redact sensitive content, and connect to custom AI services, helping automate common PDF workflows.

Create Excel and PowerPoint files with JavaScript

Essential Studio 2026 Volume 3 introduces JavaScript Excel and PowerPoint Libraries (Preview) for browser and Node.js applications. These libraries allow developers to create and process Office documents programmatically without requiring Microsoft Office installations.

• JavaScript Excel Library: Create and edit workbooks with formulas, ranges, charts, tables, validation, protection, and programmatic formatting.
• JavaScript PowerPoint Library: Generate presentations programmatically with layouts, master slides, charts, tables, images, shapes, sections, and speaker notes.

Together, these libraries extend JavaScript-based document workflows for creating and processing Excel and PowerPoint files.

Build dynamic forms with Form Builder and Form Renderer

The new Form Builder (Preview) provides a visual drag-and-drop experience for designing dynamic forms. It includes 20+ controls, conditional logic, dependent fields, calculated expressions, validation, and schema generation.

Once a form is designed, Form Renderer (Preview) renders responsive forms directly from JSON schemas.

Together, Form Builder handles visual form creation while Form Renderer turns the resulting schema into a responsive runtime form. They can support onboarding processes, approval processes, surveys, and other form-driven business applications.

New UI controls across platforms

Volume 3 expands the UI library across .NET MAUI, Pure React, Blazor, and WinUI.

.NET MAUI

Volume 3 introduces five .NET MAUI controls:

• Diagram for flowcharts, process designers, visual editors, and organizational charts.
  

  

• TreeGrid for hierarchical data in a grid interface.
  

  

• Interactive Viewer for zooming, panning, and rotating content.
  

  

• Grid Splitter for resizable application layouts.
  

  

• Data Pager for navigating large datasets.
  

  

Pure React

Pure React gains new File Upload, Sidebar, and TreeView components for file processing, navigation, and hierarchical data, respectively.

Blazor Sunburst Chart

The new Blazor Sunburst Chart(Preview) visualizes multilevel hierarchical data through concentric rings, with drill-down, selection, highlighting, theming, and tooltip support.

WinUI Docking Manager

The new WinUI Docking Manager (Preview) supports docked, floating, tabbed, and auto-hidden windows for building flexible desktop workspaces.

Key enhancements across existing components

Beyond the new controls and libraries, 2026 Volume 3 includes enhancements across existing web, Blazor, .NET MAUI, document, and desktop products.

Here are some notable updates.

Web components [React, Angular, Vue, JavaScript, ASP.NET Core, ASP.NET MVC]

• Charts: Cartesian Charts add smart data labels to reduce overlaps, while Stock Charts improve live updates, navigation synchronization, and streaming behavior.
• DataGrid: The DataGrid adds formula cells, advanced filtering, and automatic row numbering, extending spreadsheet-like interaction for business data.
• Gantt Chart: The Gantt Chart gains flexible scheduling, drag-to-create tasks, dependency restrictions, synchronized numbering, and improved hierarchical selection.
• TreeGrid: TreeGrid adds DOM virtualization for large hierarchical datasets along with enhanced parent-child selection modes.
  

  

• File Manager: File Manager adds more control over folder behavior during drag-and-drop operations.
• Modern Rich Text Editor: The Modern Rich Text Editor delivers a modern, extensible WYSIWYG experience for structured content, with ProseMirror providing the underlying document model and editing engine.
• Headless Editor: The Headless Editor is a framework-agnostic editing foundation built on ProseMirror for modern rich-text and structured editing. It provides a modular architecture with extensible commands, extensions, schemas, NodeViews, events, and reusable editing services.

Blazor components

Blazor updates span AI experiences, data interaction, project planning, and barcode support.

• Inline AI Assist is now production-ready, enabling contextual AI-assisted interactions within Blazor applications.
• Barcode Generator adds GS1 Linear, GS1 QR Code, GS1 DataMatrix, and DotCode support.
• Gantt Chart adds task constraints, multiple event markers, weekend controls, configurable view dates, and MemoryStream export.
• DataGrid adds Excel-style editing, undo and redo, find and highlight, event-driven data binding, and custom printing.

.NET MAUI Controls

Existing .NET MAUI controls also receive several notable upgrades.

• Rich Text Editor is now production-ready.
• Cartesian Charts add gradient line series, improved trackball interactions, annotation layering, and axis styling.
• Scheduler adds hierarchical resources, appointment-overlap restrictions, month agenda view, and ICS import and export.
  

  

Document solution enhancements

Document SDK

The 2026 Volume 3 release extends document generation, conversion, extraction, rendering, and security across the Document SDK.

• JavaScript PDF Library adds PDF encryption, signature validation, text search, image markers, and annotation enhancements.
  

  

• .NET PDF Library adds SVG-to-PDF conversion along with updated Blink binaries.
• .NET Excel, Word, and PowerPoint Libraries gain improvements across custom fonts, SVG support, Markdown conversion, rendering, signatures, and validation.
• .NET Data Extraction Library adds PDF-to-Word and PDF-to-HTML conversion along with barcode and QR-code recognition.

PDF Viewer SDK: Collaboration, comparison, and AI

The PDF Viewer SDK receives updates across Web, Blazor, and .NET MAUI.

Highlights include collaborative editing, semantic PDF comparison, AI-powered form-field processing, digital signatures, annotation improvements, thumbnail navigation, and document modification tracking.

• Web PDF Viewer: The Web PDF Viewer adds real-time collaboration and semantic text comparisons for identifying document changes.
  

  

• Blazor PDF Viewer: The Blazor PDF Viewer adds AI-powered form-field detection and data extraction for scanned and image-based PDFs.
• .NET MAUI PDF Viewer: The .NET MAUI PDF Viewer adds digital signatures, thumbnail navigation, and document modification tracking.

Better collaborative document experiences

DOCX Editor SDK

The Web and Blazor DOCX Editor SDK improves collaborative review with author-based comment highlighting and customizable colors for tracked changes.

Spreadsheet Editor SDK

Spreadsheet Editor updates focus on collaboration, AI automation, and spreadsheet productivity.

• The Web Spreadsheet Editor adds real-time collaboration, including synchronized changes, live presence, editing indicators, and conflict handling. It also introduces rich-text cells and WebMCP-powered AI automation.
• The Blazor Spreadsheet Editor adds charts, data validation, Find and Replace, and additional customization events.
  

  

Desktop highlights

WinUI

WinUI adds multi-card Kanban drag-and-drop, configurable DataGrid and TreeGrid sorting, and usability improvements to Time Editors.

WinForms

• The WinForms Calculation Engine adds additional spreadsheet functions for analytics, statistics, aggregation, and multilingual scenarios.
• AI AssistView adds a customizable input toolbar, quick actions, and file attachments.

WPF

• Smart DataGrid adds AI-powered export, enabling users to create Excel or PDF exports using natural-language prompts.
  

  

• WPF DataGrid and Kanban also receive usability enhancements for editing, filtering, sorting, and multi-card interactions.

Which volume 3 updates should you explore?

With updates spanning several platforms and SDKs, start with the capabilities most relevant to what you’re building.

Explore Essential Studio 2026 Volume 3

Essential Studio 2026 Volume 3 expands what’s possible across AI, document workflows, dynamic forms, collaboration, and cross-platform UI development.

Whether you’re preparing enterprise content for RAG, generating Office documents with JavaScript, building adaptive interfaces, or adding collaborative document experiences, explore the features that fit your next project.

Ready to see what’s new?

Explore the complete What’s New and Release Notes documentation to find the features relevant to your project.

Have questions or feedback? Connect with us through the Syncfusion support and feedback channels.

The fog after a layoff shows up before the grief does, and nobody warns you about it. The mind that carried a whole career can’t hold a grocery list. It’s frightening. It’s also normal, and it lifts.

I’ve written about the last day in A Letter for the Night Your Job Ends. The money got its own post, Jobs Are Disappearing Quietly. This one is about the weeks in between, when your own mind feels like a road you can’t see.

One note first: I’m not a doctor or a therapist, and this isn’t medical advice. For twenty-five years, companies have called me on their worst day. I’ve watched a lot of people walk through the weeks after a job ends. This is what I’ve seen help.

What the Fog Feels Like

It’s a Wednesday morning, and the sun is out. You open an email from the benefits office and read the first line four times. On the fifth pass, you notice you’re holding the phone and haven’t read anything at all.

You walk into the kitchen and stand there. You came in for something. The fridge hums, and it doesn’t know either. You go back to the couch without it.

You open a job board to look at one listing. Forty minutes later, you’ve scrolled past two hundred of them and applied to none. You can’t remember a single title you saw.

The words are the worst part. For twenty years you explained your work without thinking about it. Now a plain word goes missing in the middle of a sentence. You wait for it, and it doesn’t come back.

You learn what your house sounds like at eleven in the morning. For years you were somewhere else at that hour. Now you know which floorboard creaks and when the mail truck comes. Everything you know how to do is still in there, the way the road is still there in fog.

The People Who Don’t Know Yet

Your folks call on Sunday, the way they always do. “How’s work?” “Busy,” you say. Afterward you stand in the hallway with the phone still warm, and you can’t believe how easily it came out.

They’ve bragged about your job to the neighbors for years. They still have the photo from your first day, badge clipped on, grinning like it was graduation. You can’t take that away from them yet. So you carry it for them too.

Your kid is thrilled to find you home on a Tuesday. “Are you sick?” “No, sweetheart, working from home today.” By afternoon there’s a chalk sun on the driveway and YOU’RE HOME! in big, uneven letters.

A school form comes home asking for each parent’s job. You leave your line blank and sign your name under the empty space. It’s the smallest thing you’ve done all week, and the heaviest.

Your partner says, “We’ll be fine,” in a voice working hard to sound normal. You both pretend not to notice the other one pretending. Two people who love each other, being careful in the same small house.

Somewhere in all of this, a quiet, awful question arrives: what if this is who I am now? What if the layoff took my mind along with the job? It didn’t. It only feels that way from inside the fog.

Why It Isn’t Weakness

A layoff moves a lot of things in one afternoon. Your money, your days, your name for yourself and the people you saw every morning all shift at once. Your mind puts its whole weight on the danger, so there isn’t much left for a benefits letter. That’s how people are built under threat, and it says nothing about your character.

Part of me wants to argue with this whole post. Bills don’t wait thirty days, and I know some of you can’t afford a slow month. The rules below aren’t about stopping. They’re about not making hard turns while you can’t see.

The Fog Line

On American roads, the solid white line along the right edge has a nickname. Drivers call it the fog line. When you can’t see the road ahead, you watch that line and let it guide you.

It’s the most boring thing on the road. It won’t tell you where you’re going. It keeps you on the pavement until you can see again, and right now that’s all you need from it.

For the first thirty days after a layoff, you need a fog line of your own. It’s a handful of small rules you set on a calm day, so the fog can’t argue with them. These are the ones I’d give a friend.

Rules for the First 30 Days

Low beams only. High beams make fog worse, because the light bounces back and blinds you. Trying to see the next five years does the same thing this month. Plan today and tomorrow, and let that be far enough.

No big turns. Don’t make big decisions you can’t undo in the first thirty days if they can wait. Don’t sell the house, move across the country or cash out a retirement account in a panic. If a decision can’t wait, borrow someone’s eyes and ask a person you trust to read it before you sign.

The same goes for the long email to your old manager. Write it if it helps, then save it as a draft. It reads differently in a month.

Write everything down. The fog makes memory slippery, so stop trusting yours for a while. Keep one notebook for all of it: calls, names, dates, deadlines and what the benefits office told you. When your mind drops something, the notebook catches it.

Read in small pieces. If a page won’t go in, don’t fight it for an hour. Read one page, then write one line about what it said. A forty-page severance packet can be ten short sittings, or one sitting with a friend beside you.

Put a fence around the job search. In fog, speed feels like safety, and it isn’t. Pick a start time and an end time for the search each day, and close it when the time’s up. I’d rather see you send five careful applications than fifty blurry ones.

Don’t pass anyone. Someone from your old team will post that they landed a new role in two weeks. Good for them. You can’t see their road either, and passing in fog is how people get hurt.

Tell one person where you are. Before a drive in bad weather, you tell someone your route. Do the same with one person this week. Give them the true sentence, the one you’ve been replacing with “busy.”

When to Talk to Someone

The fog after a layoff should thin out a little each week. It won’t lift in a straight line. You’ll have a clear Monday and a thick Tuesday, and that’s normal too.

Talk to a doctor or a counselor if it’s getting thicker instead of thinner after a few weeks. The same goes if you can’t sleep, can’t stop sleeping, or stop eating. Reach out if you need a drink to get through the afternoon, or if nothing feels worth doing anymore.

Look in your severance papers for an employee assistance program. Some still cover a few free counseling sessions after the last day. Asking for help in fog is pulling over, the thing driving manuals tell you to do when you can’t see.

If you ever think about hurting yourself, don’t wait for the fog to lift, and reach a person today. In the US, call or text 988 for the Suicide and Crisis Lifeline. Anywhere else, Find A Helpline lists free, confidential lines in your country. Someone is there to answer right now, and you’re worth the call.

The Fog Lifts

Morning fog doesn’t lift because you fight it. It burns off as the day warms up, a little at a time. Then, all at once, you can see the fields on both sides of the road.

The fog after a layoff works the same way. One day you’ll read a whole article and notice only at the end that you read it. You’ll pick lunch in ten seconds. The word you need will be right there, where it always was.

And one Sunday, your folks will call and ask how work is. You’ll tell them the truth, all of it, and there’ll be a long pause on the line. Then one of them will say, “We were never proud of the job. We were proud of you.”

You’ll cry in the hallway, and it’ll be the good kind. You didn’t lose your mind or your skill. You lost a job, and your head needed time to catch up.

The chalk sun is still on the driveway. The road was there the whole time.

Under all of this sits one question: who are you once the work is gone? I wrote about it in Who Are You When the Work Takes Seconds? It’s an essay about where your sense of self lives when the job stops holding it.

The essay is one of thirty in my book AI: Nobody’s in There. But we’re still in here. All of them are free to read online, and the paperback is on Amazon.

Go easy on yourself this month. You’d say the same thing to a friend.

The fog after a layoff is not who you are now, it is weather.

Published by Pinal Dave on SQLAuthority. More of my work at pinaldave.com.

First appeared on The Fog After a Layoff: Rules for the First 30 Days

Today, we are announcing the general availability of DiskANN Vector Index & Search across Azure SQL Database, Azure SQL Managed Instance with the always-up-to-date update policy, and SQL database in Microsoft Fabric.

DiskANN brings scalable approximate nearest-neighbor search directly to the SQL engine. Developers can store vectors alongside relational data and combine vector similarity with the filters, joins, security policies, and transactional data their applications already rely on.

Why Vector indexes matter

Applications use embeddings to represent the semantic meaning of text, images, products, documents, and other content as vectors. Vector similarity search can then find results that are conceptually related to a user’s query, even when they do not contain the same keywords.

Without a vector index, exact vector search calculates the distance between the query vector and every qualifying vector. It must then identify and order the closest matches.

This approach can work well for smaller datasets or workloads that require exact results, but its computational cost increases as the number of vectors grows.

Efficient similarity search with DiskANN

A DiskANN vector index provides an approximate nearest-neighbor search path designed to identify relevant candidates efficiently as the dataset grows.

Instead of evaluating every vector, the index builds a navigable graph so that the database engine can navigate toward the most promising candidates. This reduces the amount of work required to identify similar results while maintaining high search quality.

Why is it called DiskANN?

The way I like to remember it –  ANN means don’t search everything, and Disk means don’t keep everything in memory More specifically the name DiskANN describes two important aspects of its design:

• ANN Approximate Nearest Neighbor: The search does not compare the query vector with every vector in the dataset. Instead, it efficiently explores the index to identify likely nearest neighbors.
• Disk: DiskANN was designed so that large vector indexes don’t have to depend entirely on expensive RAM. It can use SSD-backed storage while maintaining a relatively small memory footprint, which is what helps it scale to massive datasets

From preview to production

When we introduced DiskANN vector indexes in public preview, we invited customers to begin building vector-search applications directly on their operational data. Customer feedback helped shape the experience that is generally available today:

• DiskANN-based approximate nearest-neighbor search for large vector collections
• Full INSERT, UPDATE, and DELETE support, with asynchronous vector-index maintenance
• Iterative filtering to find the requested number of qualifying results while applying relational predicates
• Optimizer-driven execution, allowing SQL to choose between exact and approximate vector-search strategies based on the query context
• Composable T-SQL, so vector search works with existing filters, joins, business logic, and security controls

The result is vector search that behaves like part of SQL not a separate retrieval system that applications must operate and reconcile with their database. Read more here

One query from development to production

Applications can express that approximate results are acceptable using TOP (N) WITH APPROXIMATE:

SELECT TOP (10) WITH APPROXIMATE
    p.ProductId,
    p.Name,
    s.distance
FROM VECTOR_SEARCH(
    TABLE = dbo.Products AS p,
    COLUMN = Embedding,
    SIMILAR_TO = @query_vector,
    METRIC = 'cosine'
) AS s
WHERE p.IsActive = 1
ORDER BY s.distance;

The same query can continue to work as the application and its data grow. The optimizer evaluates the complete query including the data size, predicates, available indexes, and requested result count and selects an appropriate execution strategy.

Applications can therefore combine similarity search with familiar relational requirements, such as:

• Returning only products currently in stock
• Searching documents the current user is authorized to access
• Restricting recommendations by region or category
• Joining vector-search results with operational tables
• Applying existing row-level security policies

This allows developers to add semantic retrieval without giving up the relational and transactional capabilities already used by their applications. Read more here

See the evolution of vector search

In a new Data Exposed episode, we trace how Azure SQL vector search evolved as customer applications moved from prototypes toward production. The episode also demonstrates vector search across one billion vectors in Azure SQL Database Hyperscale.

Watch it here 

Get started

Ready to add vector search to your application?

• • CREATE VECTOR INDEX documentation
  • VECTOR_SEARCH documentation
  • DiskANN Vector Index in SQL Quickstart
  • Azure SQL Database Vector Search Samples
  • DiskANN vector-index improvements
  • How optimizer intelligence changes vector search

The post DiskANN Vector Index and Vector Search Are Now Generally Available in Azure SQL appeared first on Azure SQL Dev Corner.