VS Code 1.142 Adds Agent Mode, AI Use Hits 31% [2026]

Microsoft shipped Visual Studio Code 1.142 on 14 October 2026, the latest in a run of releases that has turned the world’s most widely used code editor into something closer to an AI collaborator than a text-editing tool. The update lands less than two months after VS Code 1.137 introduced scheduled agent automations and voice-driven coding sessions, and it arrives at a moment when a new Stack Overflow survey shows AI agents have moved from novelty to daily habit for a meaningful slice of the developer population. For Irish software teams, many of whom build on Microsoft’s stack or ship through Azure DevOps pipelines, the release is less a minor point update and more a signal of where the entire code editor market is heading.

Google · Preferred Sources

Don't miss new tech stories on Google

Add TrendinTech Ireland once in the Google app and our stories appear in your news suggestions.

Add Now

What Shipped in VS Code 1.142

VS Code 1.142 followed a tight cadence of releases through September and October 2026, with 1.141 landing on 7 October and 1.140 on 30 September, preceded by 1.139 on 23 September and 1.138 on 16 September. That run of five releases in under a month is faster than the editor’s historical monthly rhythm, and Microsoft’s own VS Code Blog has been publishing release notes at the same pace. The headline work behind VS Code 1.142 is a unified inline-suggestions model that merges three previously separate systems: standard code completions, next-edit suggestions, and longer-distance suggestions that look further ahead in a file than a single line. Alongside that, the team has been refining ghost-text behaviour, the faint preview text that shows a suggested edit before a developer accepts it, aiming to make the interaction feel less intrusive during normal typing.

None of this is happening in isolation. VS Code 1.142 is the clearest evidence yet that Microsoft sees the editor’s future as agentic rather than purely assistive, a shift that puts it on a collision course with dedicated AI-first editors covered in trendintech.com’s own look at Claude Code vs Copilot vs Cursor.

From Autocomplete to Agents: The Bigger Shift Behind the Release

For most of its history, VS Code’s AI features were built around autocomplete: a model suggests the next few lines, a developer accepts or rejects them, and the loop repeats. Agent mode breaks that loop. Instead of suggesting the next line, an agent can be handed a multi-step task, read across several files, make the edits, run tests, and report back. VS Code 1.142’s unified suggestions model is effectively the connective tissue between the old autocomplete era and this newer agentic one: it lets the same underlying AI system handle a quick inline fix and a longer multi-file change through one consistent interface rather than a patchwork of separate features.

That transition mirrors a pattern already visible in the browser. Chrome’s developer tools team rolled out Chrome 153 and Chrome 154 to stable over the same window, and its October 2026 DevTools update adds agent-oriented JavaScript execution controls alongside a new Chrome DevTools MCP layer for security and memory analysis. When both the leading editor and the leading browser’s developer tooling move toward agent-shaped interfaces in the same month, it stops looking like a feature trend and starts looking like a platform shift.

Scheduled Automations and Voice Agents: What 1.137 Started

To understand why 1.142 matters, it helps to look back two releases to VS Code 1.137, published on 9 September 2026. That release, documented in Microsoft’s own 1.137 release notes, introduced scheduled automations in preview: developers can now configure an agent to run a recurring task on an hourly, daily, weekly, or on-demand basis without manually kicking it off each time. The same release added an experimental Agents window that surfaces GitHub issues and pull requests directly inside the editor, so a developer can triage incoming work without first opening the repository in a browser. Perhaps the most striking addition in 1.137 was voice interaction with agents, letting developers describe a task out loud rather than typing a prompt.

Taken together, 1.137 through 1.142 read as a single sustained push rather than a series of disconnected updates. Scheduled automations give agents a sense of time. The Agents window gives them a view into project management workflows. Voice interaction changes how a request even gets issued. And the unified suggestions model in 1.142 ties the inline, moment-to-moment coding experience back into that same agentic system. It is the kind of layered rollout that suggests a roadmap decided months in advance rather than a scramble to match a competitor’s headline feature.

The Numbers: How Many Developers Actually Use AI Agents

The release would matter less if developers were not actually using these features, and the clearest recent data on that comes from Stack Overflow’s 2025 Developer Survey. Stack Overflow found that “84% of developers use or plan to use AI tools in their development process,” according to the 2025 Stack Overflow Developer Survey. That is a broad category covering everything from basic autocomplete to full agents, so Stack Overflow also measured the narrower slice: “31% of developers currently use AI agents, while 38% do not plan to use them,” the organisation reported in its press archive summarising the survey. That gap, 84% for AI tools broadly against 31% for agents specifically, is the gap VS Code’s 1.137 to 1.142 releases are built to close.

Among the developers who have crossed that line, the reaction is largely positive. Stack Overflow reported that “among developers who have used AI agents at work, 69% agree that the agents increased their productivity.” On editor choice specifically, Stack Overflow’s data shows where VS Code sits relative to the AI-native challengers: “Visual Studio Code and Visual Studio remain the top IDEs used by developers, while Cursor is used by 18%, Claude Code by 10%, and Windsurf by 5%.” For readers comparing the AI model layer that sits behind many of these agents, trendintech.com has separately covered how Grok, ChatGPT, Claude and Gemini stack up on price, which is relevant background since several of VS Code’s agent features can be pointed at different underlying models.

VS Code Release Cadence: September to October 2026

Version	Release Date	Key Addition
1.138	16 September 2026	Incremental agent and editor fixes
1.139	23 September 2026	Incremental agent and editor fixes
1.140	30 September 2026	Incremental agent and editor fixes
1.137	9 September 2026	Scheduled automations preview, Agents window for GitHub issues and pull requests, voice interaction with agents
1.141	7 October 2026	Continued refinement of unified suggestions work
1.142	14 October 2026	Unified inline-suggestions model combining completions, next-edit, and longer-distance suggestions

The table understates how much ships in each cycle. Microsoft’s developer tools group also pushed Visual Studio Code CMake Tools 1.24 in the same window, adding the ability to run tests from the editor and save preset-backed cache edits as user overrides, a smaller but telling sign of how much parallel engineering is going into the VS Code ecosystem beyond the core editor.

Competitive Comparison: VS Code, Cursor, Windsurf and JetBrains Rider

VS Code’s agent push does not happen in a vacuum. Cursor built its entire pitch around AI-first editing and, per Stack Overflow’s survey, already holds 18% developer usage. Windsurf sits at 5%. JetBrains, meanwhile, is pushing its own agent capabilities through Rider, its .NET-focused IDE, with the JetBrains .NET Tools blog confirming Rider 2026.3 EAP builds shipping through early and mid-October 2026. The comparison below lines up what is publicly confirmed about each tool’s position.

Editor	Reported Developer Usage	Agent Capability	Latest Confirmed Version
VS Code / Visual Studio	Top IDE overall (exact combined share not disclosed by Stack Overflow)	Agent mode, scheduled automations, voice agents, unified suggestions	1.142 (14 October 2026)
Cursor	18% of developers	AI-first editing built around agent workflows	Not specified in this report
Claude Code	10% of developers	Terminal-based coding agent	Not specified in this report
Windsurf	5% of developers	AI-first editing with agent features	Not specified in this report
JetBrains Rider	Not disclosed in this report	In active development, EAP builds testing new tooling	2026.3 EAP6, build 263.6800.37 (12 October 2026)

The pattern that stands out is how little pricing or feature differentiation is left between these tools and how much the competition has shifted to agent reliability and workflow depth instead. That echoes the dynamics trendintech.com found in its comparison of local LLM tools like Ollama and LM Studio, where the real differentiator was not raw capability but how smoothly a tool fit into an existing workflow.

A Brief History: How VS Code Got Here

Visual Studio Code launched from Microsoft in 2015 as a free, open source editor built on web technologies, a sharp departure from the company’s older, heavier Visual Studio IDE. Its early pitch was speed and extensibility: a lightweight core with an extension marketplace that let developers bolt on support for almost any language or framework. That strategy paid off, and VS Code has topped Stack Overflow’s annual developer survey as the most used development environment for several years running. The addition of GitHub Copilot in 2021 marked the editor’s first serious AI integration, turning autocomplete into AI-assisted autocomplete. What has changed since is the ambition: Copilot-era VS Code suggested lines, while 1.137 through 1.142-era VS Code is being built to run tasks, watch repositories, and respond to a spoken instruction. The scaffolding Microsoft has added over the past decade, the extension model, the marketplace, the massive existing user base, is exactly what makes it plausible for VS Code to pivot into an agent platform rather than losing ground to AI-native newcomers.

Why This Matters for Ireland’s Developer Market

Ireland’s technology sector leans heavily on Microsoft’s ecosystem, from Azure-hosted enterprise software to the Windows-first desktop estates common in Irish banks, insurers, and public bodies. A free, open source editor pushing deeper into agentic workflows changes the economics of adopting AI-assisted development for companies that might balk at paying per-seat for a dedicated AI-first tool. It also matters for Ireland’s large multinational tech employer base. Firms that have recently expanded engineering headcount in Dublin, including the kind of hiring trendintech.com covered around OpenAI’s Dublin expansion, are the same employers whose engineering teams are most likely to be early adopters of agent-based coding workflows, since they already run AI-heavy product stacks internally.

There is a security angle too. Scheduled automations that run unattended on a repository, even in preview, raise the same class of questions Irish compliance and security teams have had to work through for other always-on AI agents: what permissions does the automation have, what happens if it is triggered against the wrong branch, and who is accountable if an unattended agent merges something it should not have. None of that is unique to VS Code, but the scale of VS Code’s installed base means any governance gap here has a much larger blast radius than a niche tool would.

Windows App SDK, Python 3.14 and the Wider Toolchain Shift

VS Code is not the only piece of Microsoft’s developer stack moving fast this month. Windows App SDK 2.5.1 is the current stable release as of 14 October 2026, according to Microsoft’s own developer update page, while a 2.5.4 experimental build is already testing AI compute-device selection along with updated speech-recognition and language-model APIs. The October 2026 WinUI Gallery 3.0 release is built on that experimental SDK branch and showcases upcoming controls including an experimental TableView. Microsoft also pushed its Windows App Development CLI to version 0.6.0, adding a `winapp new` command that scaffolds WinUI 3 applications from official templates, a small but practical change that cuts down the boilerplate needed to start a new native Windows app.

Outside Microsoft’s own stack, Red Hat confirmed that Python 3.14’s free-threaded build is now available in Red Hat Enterprise Linux, per its developer tools update feed. Free-threaded Python removes the long-standing Global Interpreter Lock constraint for workloads that opt in, and its arrival in an enterprise Linux distribution is a signal that production teams, not just hobbyists, are expected to start testing it. For teams experimenting with running large models locally alongside these toolchain updates, trendintech.com’s guide to running DeepSeek V4.1 Flash locally covers the hardware side of that equation.

Chrome DevTools’ Parallel Agent Push

The same week that VS Code 1.142 shipped, Chrome’s developer tools team was making its own agent-shaped bet. Chrome 153 and 154 both reached stable, and the DevTools updates that came with them include fine-grained JavaScript execution controls designed for agent-driven debugging, soft-navigation support that helps tools track single-page app route changes without a full reload, performance-tier overrides for simulating slower devices, and a new Chrome DevTools MCP layer focused on security and memory analysis. MCP, the Model Context Protocol, is the same connective layer increasingly used to let AI agents call into developer tools directly rather than through a human-operated UI, which means Chrome’s own tooling team is building toward the same agent-accessible future as Microsoft’s editor team.

Market Impact: What This Means for Enterprise Dev Teams

For engineering leaders, the practical impact of VS Code 1.142 is less about any single feature and more about where the baseline is moving. When the free, default editor for a large share of the professional developer population ships scheduled agent automations and voice-driven coding as standard features rather than paid add-ons, it puts pressure on every AI-first competitor to justify its price. Cursor, Windsurf, and similar tools built their businesses on being ahead of VS Code’s native AI features. As that gap narrows, their pitch increasingly has to rest on execution quality, speed, and reliability rather than feature availability alone.

It also changes procurement conversations inside larger organisations. A security or platform team evaluating whether to approve an AI coding agent for company-wide use now has to weigh a dedicated AI-first tool against a capability that is already arriving, piece by piece, inside the editor most of their developers already have installed and already have approved. That does not make VS Code’s agent features automatically safer or more suitable, but it does mean the cost-benefit calculation for competing tools has to clear a higher bar than it did a year ago.

Risks and Friction Points Developers Are Flagging

Not every part of this shift is friction-free. Preview features, by definition, are less stable than shipped ones, and scheduled automations running unattended against a live codebase carry more operational risk than an agent a developer actively supervises in real time. Voice interaction raises its own questions in shared office environments or open-plan settings, where dictating a coding instruction out loud is not always practical or private. And the underlying trend Stack Overflow measured, 38% of developers saying they do not plan to use AI agents at all, suggests a meaningful part of the developer population remains unconvinced regardless of how deeply agent features get built into their default editor. Adoption of a feature and trust in a feature are not the same thing, and VS Code’s rapid release cadence will need to be matched by an equally rapid cadence of fixing the rough edges that preview features inevitably surface.

Five Predictions for VS Code and AI-Assisted Coding

Scheduled automations will likely graduate from preview to a stable, generally available feature within the next several release cycles, given how central they already are to Microsoft’s roadmap.
Voice-driven coding interaction will probably spread beyond VS Code as competitors look to match a feature that lowers the barrier to issuing agent instructions.
JetBrains and other IDE makers are likely to accelerate their own agent and automation features in Rider and sibling products to avoid ceding ground to VS Code’s free pricing advantage.
Enterprise security and platform teams will probably introduce formal policies specifically governing unattended, scheduled coding agents, separate from existing AI-assistant policies.
Python’s free-threaded build reaching Red Hat Enterprise Linux is likely to accelerate testing of multi-threaded Python workloads in production environments over the next year.

What Developers Should Do Now

Teams running VS Code should treat the 1.137 through 1.142 run as a single feature set to evaluate together rather than six separate minor updates. That means testing scheduled automations against a non-critical repository before trusting them anywhere near production branches, reviewing what permissions any agent-driven GitHub integration actually has, and deciding internally whether voice-driven prompts fit the team’s working environment at all. None of these features are mandatory to use, and VS Code’s extension-based architecture means teams can continue working exactly as before if they choose to. The risk is less that these features are bad and more that they get adopted informally by individual developers before a team has agreed on how they should be governed.

How This Compares to Other Editor Tooling on trendintech.com

Readers weighing lighter-weight alternatives to a full agentic IDE may find trendintech.com’s look at mobile HTML editors like Acode and Spck useful context for how differently the editor market looks once mobile and lightweight use cases are factored in, away from the desktop-first, agent-heavy world VS Code 1.142 is building toward.

Frequently Asked Questions

What is new in VS Code 1.142?
VS Code 1.142, released 14 October 2026, introduces a unified inline-suggestions model that combines standard code completions, next-edit suggestions, and longer-distance suggestions into one system, alongside refined ghost-text behaviour.

What did VS Code 1.137 add?
VS Code 1.137, released 9 September 2026, introduced scheduled automations in preview for recurring agent tasks, an experimental Agents window for inspecting GitHub issues and pull requests, and voice interaction with coding agents.

How many developers actually use AI coding agents?
According to Stack Overflow’s 2025 Developer Survey, 31% of developers currently use AI agents, while 38% say they do not plan to use them, against a much broader 84% who use or plan to use AI tools of any kind.

Is VS Code still the most used code editor?
Yes. Stack Overflow’s data shows Visual Studio Code and Visual Studio remain the top IDEs overall, ahead of Cursor at 18% usage, Claude Code at 10%, and Windsurf at 5%.

Do AI coding agents actually improve productivity?
Stack Overflow reported that among developers who have used AI agents at work, 69% agree the agents increased their productivity.

Are VS Code’s scheduled automations safe to use in production?
They remain in preview as of the 1.142 release, so most engineering teams are advised to test them against non-critical repositories first and define permission and approval policies before allowing them near production branches.

Is VS Code free?
Yes, Visual Studio Code remains free and open source, which is part of why its agent features put pricing pressure on paid AI-first competitors like Cursor and Windsurf.

What should Irish development teams do about this release?
Review the combined 1.137 to 1.142 feature set as a single package, test scheduled automations and agent integrations in a controlled environment, and set internal governance policies for unattended coding agents before enabling them broadly.

VS Code 1.142 Ships Agent Mode as AI Use Hits 31% [2026]