Code Editors and IDEs Release Notes
Release notes for code editors and integrated development environments
Products (12)
Latest Code Editors and IDEs Updates
- Sep 2, 2026
- Date parsed from source:Sep 2, 2026
- First seen by Releasebot:Sep 3, 2026
Self-hosted machines
Cursor adds self-hosted machines for local tool execution, keeping code, builds, and secrets inside your network. It also brings dynamic machine pools, support for cloud agents on existing infrastructure, and computer use on Linux and Mac for browser and desktop control.
Cursor supports self-hosted machines, which let you keep tool execution entirely in your own network.
Your codebase, build outputs, and secrets all stay on internal machines running in your infrastructure, while the agent handles tool calls locally.
Dynamic pool scheduling
My Machines connects a single laptop or VM to your account for personal workflows.
Team pools are named queues of workers for a team or enterprise. Capacity can grow as requests arrive and shrink when workers disconnect, so your self-hosted machines can scale with demand. Pools are not tied to one repository: name the pool, and any available worker can claim the request.
Pools can also hibernate idle machines, then restore within a reconnect window when a follow-up arrives, so you don't keep expensive capacity warm just for the next prompt.
Run on your sandboxes
Cloud agents can now execute on infrastructure you already use, including from AWS Lambda, Coder, Cloudflare, Daytona, Modal, Namespace, Vercel, and E2B.
Computer use on Linux and Mac
Self-hosted workers now support computer use on Linux and Mac. With the right desktop packages, an agent can click, type, take screenshots, and drive the browser. You can watch its desktop or take control from Cursor.
Original source - Sep 2, 2026
- Date parsed from source:Sep 2, 2026
- First seen by Releasebot:Sep 2, 2026
1.18.0
Zed releases a broad update with refreshed Markdown preview styling, inline debugger values for C and C++, and persistent File Finder history, while also improving AI models, Git workflows, terminal input, Linux performance, and fixing many editor, panel, and stability issues.
This week's release includes refreshed Markdown preview styling, inline debugger values for C and C++, and File Finder history that persists across workspace sessions.
Shipped by the Zed Guild đĄď¸
- Fixed the inline assistant doing nothing when the model emitted reasoning before its tool call. (#61220; thanks aviatesk)
- Fixed some Pyright and BasedPyright settings not being applied properly. (#62673; thanks lingyaochu)
- Fixed bookmarks not disappearing from the gutter immediately after clearing all bookmarks. (#62847; thanks austincummings)
- Fixed the LSP status tooltip to display the target script path for runtime-managed language servers. (#62919; thanks lingyaochu)
Learn about the Zed Guild.
Features
AI
- Added support for GPT-5.6's 1M-token context window on Amazon Bedrock. (#63069; thanks 5herlocked)
- Added Gemini 3.5 Flash-Lite to the Google AI models. (#63129; thanks ArneshBanerjee)
- Added Grok 4.5 and Grok 4.6 to the xAI models. (#63234)
- Added the ability to reload a broken external agent connection from the Agent Panel without restarting Zed. (#62669; thanks gaojunran)
- Added the configurable edit_predictions..prediction_debounce setting for inline completions. (#61568; thanks ddeityy)
- Added settings autocomplete for the OpenAI subscription language model provider. (#62989)
- Improved memory usage during long agent sessions. (#63053; thanks Prohect)
- Improved support for Claude Fable 5.1. (#63571)
- Improved language model connection errors by identifying the unreachable host. (#62990)
Git
- Added permalink support for Tangled repositories. (#63051)
- Added the git: toggle diff base action to toggle the diff base between HEAD and the default branch. (#62966)
- Improved the Git Graph and Git history views to initially focus the search editor. (#63002)
- Improved the Git Panel's âDiscard Tracked Changesâ option to affect only files within the directory where its context menu was opened. (#62872)
Debugger
- Added inline values for C and C++ when debugging with CodeLLDB or GDB. (#46705; thanks xcb3d)
- Added separate Continue Program and Continue Thread controls for debug adapters that support single-thread execution. (#63135)
Languages
- Added support for configuring Bash Language Server through lsp.bash-language-server.settings. (#57487; thanks CathalMullan)
- Improved Markdown preview styling with better typography, spacing, and rounded inline code backgrounds. (#63118; thanks archcorsair)
- Improved the tabular data preview by moving its settings into a popover, freeing more vertical space for the table. (#62812; thanks HalavicH)
Terminal
- Added terminal support for ctrl-alt- keystrokes. (#62891; thanks nbdd0121)
Linux
- Improved rendering performance by prewarming font match caches. (#63158)
- Improved rasterization performance by caching glyphs after measuring bounds. (#63209)
Other
- Added a Clear button to the Settings window's search field that clears the search and resets the navigation list. (#56033; thanks asgoode)
- Added the in_preview keybinding context for editors in preview mode. (#61777; thanks krawitzzZ)
- Improved file finder history to include recently opened files from previous sessions. (#55034; thanks anilzeybek)
- Improved typing and rendering performance in projects that use .editorconfig files. (#63323)
- Improved text input in GPUI web applications on Android by adding support for Gboard. (#63201)
- Improved responsiveness when closing projects that are loading large text files. (#62831; thanks 39ali)
Bug Fixes
- Fixed a bug where last-session workspace state was not preserved or restored after closing the last macOS window with the red close button when on_last_window_closed was set to "quit_app" and restore_on_startup was set to "last_session". (#60477; thanks Caleul)
- Fixed a bug where the terminal panel failed to open in some circumstances and panels switched back to the restored panel during startup. (#62712)
- Fixed OAuth authentication failing with MCP servers that require non-default scopes. (#63370; thanks blue42u)
- Fixed a bug where Zed windows sometimes remained blank when first opened on X11. (#62081; thanks mTvare6)
- Fixed Continue Program resuming only the selected thread with some debug adapters instead of all threads. (#63135)
- Fixed global gitignore entries incorrectly matching directories outside the opened project and marking the entire project as ignored. (#62130; thanks floesche)
- Fixed language model provider rejections being retried and displayed without their actionable provider message. (#63095)
- Fixed performance issues caused by overly large LSP hovers. (#63022)
- Fixed the UI freezing in fullscreen on some Wayland compositors. (#60690)
- Fixed a bug where edits in GPUI web applications could be applied to stale text. (#63201)
- Fixed a bug where keybindings in the Git Panel context menu changed after the first frame. (#63021)
- Fixed a bug where the Debug Panel did not immediately reflect toggled breakpoints. (#63240; thanks caseyavila)
- Fixed a crash when an OpenAI-compatible model sent a usage-only stream chunk with null cache token fields. (#62983; thanks enoch-robinson)
- Fixed a crash when dispatching profiled GPUI actions in web applications. (#63183; thanks labcoder)
- Fixed a panic when closing a window under X11. (#63089)
- Fixed completion menu labels overlapping documentation text for long completion items. (#63581; thanks itsfuad)
- Fixed CSV previews not appearing when no project was open. (#63350)
- Fixed a bug where file names containing control characters, such as tabs and newlines, rendered unreadably in tabs, the Project Panel, the file finder, the tab switcher, and terminal tab titles. (#62875; thanks Taqinou)
- Fixed focus being stolen when opening modals. (#63512; thanks mateioprea)
- Fixed a bug where the Agent Panel occasionally forgot the selected agent. (#61147; thanks tnayuki)
- Fixed a bug where the Agent Panel selected an uninstalled external agent when opening a project or restoring an unsent draft, which prevented the thread from starting. (#62497; thanks white-hat)
- Fixed a bug where the agent edits panel and review diff showed whole-file overwrites as entirely added lines with no removals. (#62267; thanks drawing)
- Fixed a bug where the file finder jumped to the wrong file when the mouse hovered over another entry after pressing cmd-p. (#61716; thanks HuaGu-Dragon)
- Fixed a bug where the Python run button did not appear in the gutter for if name in ("main", ...) -style main guards. (#58911; thanks tlium)
- Fixed a bug where the X11 urgency hint was never cleared, which left Zed showing a permanent attention indicator in taskbars and docks after the first notification. (#61619; thanks zquestz)
- Fixed Anthropic billing, conflict, and timeout errors being handled incorrectly. (#62984)
- Fixed duplicate file entries in the Project Panel on macOS when language servers returned definition paths whose casing differed from the worktree root. (#61392; thanks NeverDieOne)
- Fixed environment variables defined in tasks.json not being applied when debugging Python tasks. (#63154; thanks andreyfedoseev)
- Fixed extension tags obscuring action buttons on extension cards. (#63153; thanks joshkent94)
- Fixed function bodies not folding when their arguments spanned multiple lines. (#62245; thanks anxyum)
- Fixed handling of OpenAI overload errors. (#63238)
- Fixed idle Wayland windows waking at the display's refresh rate. (#60690)
- Fixed labels in Mermaid diagrams not soft-wrapping correctly. (#62931)
- Fixed localized Windows Explorer context menu titles falling back to English. (#60634; thanks justjavac)
- Fixed new windows opening partially off-screen on secondary Windows monitors with different DPI scaling. (#62859; thanks dineshyadav03)
- Fixed opening remote paths from SCP-style SSH URLs with IPv6 hosts. (#62157; thanks cppcoffee)
- Fixed overlapping range-formatting results duplicating text. (#62935)
- Fixed read-only files being formatted or saved when format_on_save was enabled. (#62921; thanks mateioprea)
- Fixed same-line regex replacements using lookahead or lookbehind. (#61900; thanks IbrahimKhan12)
- Fixed streaming failing for Mistral-hosted models that omitted the object field in stream chunks, such as GLM 5.2. (#62517; thanks ummon-v)
- Fixed terminal support for the alt-f5 keystroke. (#62891; thanks nbdd0121)
- Fixed text being duplicated after moving the caret during composition in GPUI web applications. (#63201)
- Fixed text wrapping so closing punctuation and slashes never began a wrapped line. (#62743; thanks alii)
- Fixed the lsp_results_location setting being ignored by the editor: go to declaration and editor: go to type definition actions. (#61060; thanks xdBronch)
- Fixed the âDiscard Tracked Changesâ option being enabled for files in the Git Panel's âUnstagedâ section. (#62872)
- Fixed the terminal panel dismissing an open modal when terminals finished restoring during startup. (#63097; thanks butvinm)
- Fixed undoing a file rename leaving behind directories created by the rename. (#60082)
- Fixed window-manager rules and taskbar grouping failing to recognize Zed under X11 and XWayland. (#62848; thanks HundredBillion)
- Fixed Windows artifacts around color emoji glyphs. (#63225; thanks 39ali)
- Fixed Zed overriding window-manager focus policies when activating windows. (#62081; thanks mTvare6)
- Fixed macOS showing one-time code AutoFill suggestions in regular Zed text inputs. (#60116)
Breaking Changes and Notices
- Improved debugger step keybindings across platforms to match VS Code defaults: f10 now steps over and f11 steps into while execution is paused, while fullscreen remains available outside paused sessions. (#58729; thanks Andrej730)
- Removed Gemini 2.5 Flash-Lite, Gemini 2.5 Flash, Gemini 2.5 Pro, and Gemini 3.1 Flash-Lite from Google AI, and made Gemini 3.5 Flash-Lite the default and fast model. (#63131)
- Removed GLib libraries from Linux release archives to prevent conflicts with system plugins. (#61593)
All of your release notes in one feed
Join Releasebot and get updates from Cursor and hundreds of other software products.
- Sep 2, 2026
- Date parsed from source:Sep 2, 2026
- First seen by Releasebot:Sep 2, 2026
Run cloud agents on machines you manage
Cursor adds Self-Hosted Machines for cloud agents, letting teams run agents on their own infrastructure while Cursor still manages planning and orchestration. It also expands pool scaling, sandbox provider support, and Linux computer use for more flexible, secure agent workflows.
Cursor cloud agents can execute on dynamically scheduled pools of machines inside your network. You manage the underlying infrastructure, while agents are still started and managed from Cursor.
This gives teams more control over where agents execute and what infrastructure they use. Agents can work next to internal services and source control, run on custom hardware, or use operating systems and build pipelines that are difficult to package as a Cloud Agent build.
Cloud agents now create more than 60% of the pull requests we merge internally and are taking on a growing share of software work at many of the largest enterprises we work with. As their role expands, the machines they run on matter more too. These new capabilities make it practical for teams to provide and manage that infrastructure at scale.
With Lambda MicroVMs as the compute layer for Cursor Cloud Agents, developers can run AI-powered coding agents in their own AWS account. Each machine launches near-instantly from a snapshot, suspends when idle, and resumes with full state. Your coding agents benefit from Lambda's fast startup, strong isolation, and zero fleet management, while Cursor orchestrates the work.
Control where agents execute
Cursor-hosted environments remain the default for cloud agents. Each session runs on a dedicated VM inside the Cursor cloud, with its dependencies installed and its own network controls. Per-agent isolation, secret redaction, egress controls, and signed commits meet the security requirements of most teams.
Teams generally use Self-Hosted Machines when:
- Agent tool execution needs to happen inside their network, with direct access to source control, internal services, and code repositories.
- Agents require custom hardware, such as GPUs or Macs for iOS development, or infrastructure such as Kubernetes, sandboxes, or managed VMs.
- Their operating system or build pipeline is difficult to package as a Cloud Agent build.
With Self-Hosted Machines, only the execution environment moves while the agent loop, inference, and planning remain in the Cursor cloud. Tool outputs flow back to Cursor for inference and may contain code, and agent transcripts may be processed and stored by Cursor. Teams can continue to access cloud agents from the desktop app, cursor.com, mobile, Slack, GitHub, and Linear.
Workers connect your infrastructure to the Cursor agent loop
With Self-Hosted Machines, tool execution moves from a Cursor-hosted VM to a machine in your environment. That machine holds the working copy of the repository, edits files, and runs commands. A worker connects it to the rest of the agent system.
To register a machine, run a worker by installing the Cursor CLI and running
agent worker startThis opens a long-lived outbound HTTPS connection to the Cursor cloud. When a session begins, Cursor's agent harness handles inference and planning, then sends tool calls to a dedicated worker for execution. The worker returns the results for the next round of inference. Cursor never initiates a connection into your network.
Workers can be configured in two ways.
- My Machines. This configuration connects a single laptop or VM to your account and is best suited to personal workflows.
- Pools. A pool is a named queue of workers that can serve a team or enterprise. Capacity increases as requests arrive and decreases after workers disconnect, letting your existing cloud infrastructure scale with developer demand.
Developers should have the flexibility to run coding agents on the platform that best supports their workflow, and companies should not have to compromise on control of where agents run and what they can access. The future of development will be built on powerful agents, running in secure, isolated environments.
Cloud agents adapt to your infrastructure
Worker pools can now scale in response to queued requests and serve work from any repository. We have also added support for several sandbox providers and computer use on Linux alongside Mac.
Pools scale with demand and serve any repository
Demand for cloud agents often arrives in bursts and Self-Hosted Machines pools adjust to those bursts automatically. This happens through a controller which watches the request queue and uses a spawn script supplied by the team to start machines as needed.
If a pool has an available worker, that worker claims the request. Otherwise, the request waits until more capacity becomes available, so teams do not have to decide how many machines to leave running.
Teams can set an idle timeout for each worker connection. Once it expires, the machine can reset and re-enter the pool. Teams can also preserve its workspace in case the agent receives a follow-up.
Self-Hosted Machines put teams in control of where Cursor agents run, and Vercel Sandbox makes it effortless. Every task gets an isolated sandbox on demand, no fleet to manage, and nothing sitting idle.
Leaving a machine running while its agent is idle can be expensive. But if the machine is released, the agent may need several minutes to reconstruct its workspace when a follow-up arrives. With hibernation, teams can snapshot and stop an idle machine instead. If a follow-up arrives within the reconnect window, the snapshot is restored and a worker starts with the same ID. Otherwise, the request can move to a new machine.
Pools are not tied to individual repositories. A request only needs to identify the pool, and any available worker can claim it. This lets one pool serve many repositories.
Workers run across supported sandbox providers
Self-Hosted Machines does not require building a custom sandbox layer from scratch. We partner with AWS Lambda, Cloudflare, Coder, Daytona, E2B, Modal, Namespace, and Vercel, allowing workers to be started and orchestrated wherever a team's sandboxes already run.
Cursor Self-Hosted Machines on Modal gives each Cloud Agent session a Modal Sandbox, so you can hand it a machine tailor-made for its task.
Agents control browsers on Linux and Mac
Linux workers now support computer use alongside Macs. With the required computer use dependencies installed, including Chrome or Chromium, an agent can click, take screenshots, and control the browser. You can watch its desktop or take control directly from Cursor.
You can't build iOS or macOS apps without a Mac. Namespace Devboxes spin up a real Mac for each Cursor Cloud Agent, which can now perform that work on Apple silicon.
Bring cloud agents into your environment
Teams have spent years shaping their infrastructure around how they build software. Self-Hosted Machines lets cloud agents fit more naturally into it, and we're excited to see how far teams take them.
To connect a machine or configure a pool, get started in the docs.
Original source - Sep 2, 2026
- Date parsed from source:Sep 2, 2026
- First seen by Releasebot:Sep 2, 2026
Visual Studio Code by Microsoft
Visual Studio Code 1.136
Visual Studio Code releases 1.136 with agent-focused updates for finishing pull requests, managing multi-root workspace sessions, organizing related chats, and personalizing the Agents window. It also brings improved editor, browser, terminal, and accessibility experiences.
Welcome to the 1.136 release of Visual Studio Code. This release helps you finish pull requests with agents and manage agent work across complex workspaces and related chats.
- Agent Merge (Preview): Resolve review feedback, failed checks, and merge conflicts until your pull request is ready to merge.
- Multi-root workspaces (Experimental): Work with Copilot and Claude agent sessions across all folders in a multi-root workspace.
- Chat backgrounds (Experimental): Personalize the Agents window with built-in patterns or your own images.
- Chat sessions: Organize related chats in a session hierarchy and quickly see which ones need your attention.
Happy Coding!
VS Code is rolling out gradually to all users. Use Check for Updates in VS Code to get the latest version immediately.
To try new features as soon as possible, download the nightly Insiders build, which includes the latest updates as soon as they are available.
The Story of VS Code: World Premiere
Discover the story behind VS Code, from its early beginnings to the platform millions of developers use today, and the community that helped shape it along the way.
Premiere: September 4 at 8:00 AM PT. Join us!
Agents
Multi-root workspaces in editor window (Experimental)
Settings: chat.agentHost.copilotAgent.multiRootEnabled, chat.agentHost.claudeAgent.multiRootEnabled
Copilot and Claude agent sessions in the editor window Chat view support multi-root workspaces.
This capability is currently scoped to the editor window.
Agent hooks remain scoped to a single workspace folder. If hooks are detected in multiple folders, VS Code prompts you to select the primary folder from which to load them.Redesigned new-session input
Start delegated work with less setup in the redesigned new-session input. The updated experience brings the prompt, model selection, workspace selection, and other session controls together in one layout.
Improved workspace resolution
Agents can resolve workspaces by their project name in addition to absolute paths and workspace URIs. Session tools also preserve project URIs and all working directories for multi-root workspaces.
As a result, you can make requests such as "run this in the vscode workspace" without providing a full path. If multiple workspaces have the same name, the agent reports the possible matches instead of silently choosing one. Remote workspace URIs are supported as well.
Navigate related chats and sessions
Keep related agent work organized and move between sessions and chats from the sessions list in the Agents window. Chats appear as children of their parent session, so you can understand which chats belong together without managing a collection of unrelated sessions.
Each chat row shows its own title, status, and pending approvals, so you can see which chat needs your input. Expand or collapse the hierarchy, and open, rename, move, or delete individual chats directly from the tree.
When an agent delegates independent work to multiple chats, each created chat gets a meaningful title and appears in this hierarchy.The new session or chat is placed close to its source, and the receiving request includes a source link such as Sent by another session or Sent from another chat. Select the link to return to the exact session or chat that initiated it.
Readable breadcrumbs for session files
Files created in the internal session state directory previously showed internal session identifiers in editor breadcrumbs. Breadcrumbs now use stable provider and session labels, making it easier to identify the file location without exposing implementation details.
Chat backgrounds in the Agents window (Experimental)
Settings: chat.agentSessions.preferredDarkBackgroundImage, chat.agentSessions.preferredLightBackgroundImage, chat.agentSessions.backgroundImageLayout
Personalize the Agents window with a decorative chat background: either a theme-aware pattern of built-in VS Code icons or an image of your own.
Run Chat: Set Background... to choose between the Codicons pattern and an image from your machine. The five images you used most recently are offered under Recently Used, and that list is kept on this machine only.
When you bring your own image, run Chat: Change Background Layout... to place it. Eleven layouts are available: Repeat, Stretch, Center, and each edge and corner. Moving through the list previews each layout in place, so you can judge the result before committing to it. Chat: Clear Background returns to the plain surface.
Dark and light themes keep separate backgrounds. Every dark theme shares chat.agentSessions.preferredDarkBackgroundImage and every light theme shares chat.agentSessions.preferredLightBackgroundImage, so switching between a dark and a light theme swaps the background along with it. High contrast themes suppress backgrounds entirely, and the three commands are unavailable there.
Chat content carries its own fill so that it stays readable over whatever sits behind it. Your request remains fully opaque, agent responses fade out through the padding on either side, and wide content such as Markdown tables and terminal output stays fully backed.
Learn more about personalizing chat.Agent session notifications
Settings: chat.notifyWindowOnConfirmation, chat.notifyWindowOnResponseReceived
VS Code can notify you when an agent session needs your input or finishes its work. This is especially useful when you delegate work across multiple sessions, workspaces, or VS Code windows.
By default, notifications only appear when the VS Code window is not focused. You can configure notifications separately for sessions that need input and sessions that have received a response. Notifications include a direct link back to the relevant session, so selecting one focuses the correct window and opens the session that needs attention.Agent Merge (Preview)
Setting: chat.agentMerge.enabled
Agent Merge helps you take a pull request across the finish line. It asks an agent to address review feedback, fix failed checks and merge conflicts, and rerun workflows. Agent Merge repeats this process until the pull request is ready to merge.
To try Agent Merge, enable chat.agentMerge.enabled. You can currently enable Agent Merge for a session only from the Agents window. Run Enable Agent Merge for Active Session or select the Agent Merge button in the title bar. Learn more about using Agent Merge.Agent host
The agent host lets you connect to the same agent session from multiple VS Code windows. It runs agent harnesses in a dedicated process based on the Agent Host Protocol (AHP). The agent host's Copilot agent is powered by the Copilot SDK, which aligns the agent's behavior and functionality with the Copilot CLI, the standalone GitHub Copilot app, and other Copilot products.
We're actively developing the agent host. The following screenshot shows the Copilot harness selected for an agent host in an editor window:
Learn more in the agent host documentation and our new agent host blog post, where we share why we built the agent host, what it enables in VS Code, how the architecture and open protocol work, and workflows you can try yourself.
If you have any feedback or requests, please let us know by filing an issue.Chat
Enterprise controls for dictation data
Administrators can manage the dictation model and language-model transcript cleanup through enterprise policies.
The new controls make it possible to require on-device transcription and disable language-model cleanup, so dictation remains available while preventing dictation data from being sent to cloud transcription or a Copilot model. Learn more about enterprise controls for dictation data.Accessibility
Screen Reader Optimized badge in the Agents window
The Screen Reader Optimized badge in the Agents window title bar makes the active accessibility mode easier to identify. The badge appears when Screen Reader Optimized mode is enabled. Select the badge to disable this mode.
Editor experience
Layout density for the editor window (Experimental)
Settings: workbench.experimental.modernUI, window.density.layout
Fit more content in the editor window with the Compact layout density. When workbench.experimental.modernUI is enabled, you can choose between two layout densities:
- Default layout density is the same as the current editor window layout.
- Compact layout density removes the spacing between panels and reduces inner panel spacing.
This setting is available in the Layout Density section of the Settings menu and can also be set using window.density.layout.
Code editing
Word wrapping improvements
Injected text no longer pushes wrapped lines outside the editor viewport. Word wrapping takes the visual width of color decorators, inlay hint spacing, inline progress indicators, and breakpoint placeholders into account. Please have a look below for the before and after screenshots. On the picture above you can see that rgba is slightly cropped. While on the second picture it isn't.
Integrated browser
Spell check suggestions
Right-click a misspelled word in an editable field to select a suggested correction. In sessions that use persistent data storage, you can also select Add to Dictionary.
Terminal
Reduced delay when running commands
Terminal commands executed by extensions no longer incur an unnecessary delay when shell integration is ready under certain timing conditions. Users of the JavaScript debugger who hit this case will no longer experience a five-second delay when starting their programs.
Deprecated features and settings
None
Thank you
Contributions to vscode:
- @abmahdy (Ahmed Mahdy): Bound concurrency of prompt file discovery reads PR #331855
- @DanTup (Danny Tuppeny): Note that InlayHints with the same position are shown in-order PR #175525
- @davidbitton (David B. Bitton): Add spelling suggestions to the Integrated Browser context menu PR #333043
- @JeffreyCA: Update Fig spec for Azure Developer CLI (azd) PR #331727
- @juliagongms (Julia Gong): nes: apply supportsUnifiedCompletions on the default provider path PR #332802
- @koubaki: Update error message in inlineChatIntent.ts PR #329157
- @ktsoator (ktsoator): Hover: restore scrolling for long content PR #331439
- @na2co3-ftw (na2co3): Modern UI: Fix CSS specificity for tab action fading PR #332103
- @preitinger (Peter Reitinger): Update snippet.md PR #231790
- @remcohaszing (Remco Haszing):
- Normalize end of line cursor move operation PR #296712
- Fix out of bounds text selection with line wrapping PR #262910
- Round custom line heights PR #298421
- Expose the score function in Monaco editor PR #322959
- @SimonSiefke (Simon Siefke):
- fix: memory leak in explorer viewer PR #332332
- fix: memory leak in LSP terminal completions PR #332173
- @TheNotary: Update error message for unsupported skill attributes PR #328318
- @tisilent (xiejialong): update label and description. PR #219949
- @unsupportedpastels (Mark S.): Support reasoning effort in custom agent files PR #329263
- @weidehai (io): error correction instructions PR #249715
Contributions to vscode-emmet-helper:
- @danfiedler-msft (Dan Fiedler): Pin GitHub Actions to full-length commit SHAs PR #108
Contributions to vscode-livepreview:
- @danfiedler-msft (Dan Fiedler): Pin GitHub Actions to full-length commit SHAs PR #854
Contributions to docfind:
- @danfiedler-msft (Dan Fiedler): Pin GitHub Actions to full-length commit SHAs PR #62
Contributions to node-pty:
- @danfiedler-msft (Dan Fiedler): Pin GitHub Actions to full-length commit SHAs PR #958
Issue tracking
Contributions to our issue tracking:
- @gjsjohnmurray (John Murray)
- @saroasid-web (Saswwo)
- @zotabee (zotabee)
- @lppedd (Edoardo Luppi)
- @sandstrom (sandstrom)
We really appreciate people trying our new features as soon as they are ready, so check back here often and learn what's new.
If you'd like to read release notes for previous VS Code versions, go to Updates on code.visualstudio.com.
Original source - Sep 2, 2026
- Date parsed from source:Sep 2, 2026
- First seen by Releasebot:Mar 14, 2026
- Modified by Releasebot:Sep 2, 2026
cmux by manaflow-ai
Nightly
cmux publishes a nightly universal app build with updated appcast feeds and macOS download.
Automated nightly build for 60d06c2.
cmux NIGHTLY is published as a universal app:
bundle ID com.cmuxterm.app.nightly
feed appcast.xml
compatibility feed appcast-universal.xml for older universal nightlies
Download cmux-nightly-macos.dmg
Original source - Aug 27, 2026
- Date parsed from source:Aug 27, 2026
- First seen by Releasebot:Aug 28, 2026
2026.08.27 (v0.2026.08.26.17.59)
Warp launches Warp Factories in Early Access and adds a wide set of improvements, bug fixes, and cloud agent reliability updates, including better Windows and Linux input, richer usage cost breakdowns, smarter completions, and smoother web, terminal, and transcript experiences.
New features
- Warp Factories is now live in Early Access. Read the Warp Factories launch announcement for details.
Improvements
- Removed unnecessary first-time-user banners and prompts (credits, notifications, desktop download) when viewing a shared session or conversation transcript on the web. (#15515)
- Disabled team/workspace members are now shown grayed out with an explanatory tooltip instead of appearing active. (#15258)
- Renamed the âOz Cloud API Keysâ settings page to âAPI keysâ under Cloud platform. (#15424)
- Enabled IME marked text (composition) input on Windows and Linux, previously macOS-only. (#15395)
- Added a setting to make bare right-click paste from the clipboard, with Shift+right-click still opening the context menu. (#15365)
- Added a setting to stop â#â at the start of terminal input from opening AI Command Search, so shell comments arenât interrupted. (#15340)
- Usage displays (agent block usage, conversation footer, usage history, and task details) now show a per-category dollar cost breakdown alongside credits. (#15148)
- Added a âPasteâ item to the terminal block listâs right-click context menu. (#15346)
- The markdown viewer now preserves your scroll position when switching between Rendered and Raw view modes. (#13967) â @slash9494 â¨
- Holding a modifier now shows each tabâs switch-to-tab keyboard shortcut on the tab itself. (#15221) â @JJack27 â¨
Bug fixes
- Fixed bash sessions never reporting detected shell plugins (e.g. Starship) back to Warp. (#15518)
- Fixed mismatched Shift/Command/Option key icons in keybinding chips (Command Palette, context menus, tooltips) on the web build. (#15569)
- [Windows] Fixed Warp becoming unresponsive after disconnecting and reconnecting a Remote Desktop session. (#15566) â @bbbiggest â¨
- Fixed the Teams settings confirmation dialogs (remove member, transfer ownership, leave/delete team) landing at the top or bottom edge of the window depending on the pageâs scroll position instead of staying centered. (#15542)
- Fixed automatic CLI plugin installation failing in Windows PowerShell 5.1. (#15544)
- Fixed a multi-gigabyte memory spike when editing very large files, caused by unbounded parallel text-layout work. (#15128)
- Fixed cloud agent runs occasionally getting stuck showing in progress when a child agent finished right as the parent run began tearing down. (#15432)
- Pressing Ctrl-C in an attached third-party agent harness session (e.g. Claude Code) now correctly resolves it to cancelled instead of leaving it stuck showing in progress. (#15506)
- Cloud agent runs now always report a terminal status (succeeded, failed, or error) when the agent process exits, instead of occasionally being left stuck in progress after sandbox deadlines, SIGTERM shutdowns, or exit races. (#15387)
- Fixed a memory spike when opening large files in the code editor caused by an unnecessary full-file clone. (#13508)
- Fixed completions for many commandsâ multi-argument options that were mis-modelled (e.g. git branch/git switch, rsync --info/--debug, rustc -L, dart format, hugo new, date), so they no longer offer the wrong arguments. (#15487)
- Fixed ln -s and ln -F/-f/-i not offering file path completions. (#15480)
- Fixed shell command completions resolving the wrong argument for options that take multiple arguments of differing types (such as cargo fmt -- --print-config and jq --rawfile), where a value before the last received a later argumentâs suggestions. (#15475)
- Fixed PowerShell sessions using PSReadLineâs vi edit mode submitting mangled commands. (#15476)
- Cloud agent environments no longer hang with runaway âcloningâ output when running git log, git blame, or other path-limited history commands. (#15281)
- Fixed several latent shell-integration bugs, including a crash from slicing spans on non-UTF-8 character boundaries. (#15428)
- Fixed the usage footerâs conversation-total dollar amount reading higher than the per-response breakdown by aligning both with the same charged-usage figures. (#15423)
- Fixed the agent grep tool returning a parse error, or in rare cases a wrong file/line, when a matched file path contained a colon (e.g. Windows paths, some Go module paths). (#15422)
- Retry transient server errors when resolving a managed MCP serverâs connection, instead of permanently failing the agent run on the first blip. (#15426)
- Fixed a hidden Alt+1 (Ctrl+1 on Mac) binding for Project Explorer that could not be remapped or disabled from Settings > Keyboard shortcuts. (#15310)
- Fixed stdio MCP servers failing to start on a fresh oz agent run with âUnknown PATH when trying to launch MCP command.â (#15396)
- Fixed zsh tab completions losing their descriptions for commands using clustered short options. (#15313)
- Fixed unbounded memory growth from rebuilding the launch-config search index on every rapid configuration-file change. (#15119)
- Tab autocomplete now follows symlinks to directories in WSL sessions. (#14755)
- Improved scrolling performance in transcripts with many Agent Mode blocks. (#15280)
- Fixed a âdouble cursorâ issue where a finished background block kept showing a cursor alongside the input editor or the next block. (#15322)
- Fixed a stray â(1)â badge appearing on the Code Indexing and Profiles settings sub-pages during search. (#15319)
- Fixed cloud agent runs having too small a recovery budget to retry a mid-stream connection reset, sometimes failing runs instead of automatically retrying. (#15293)
Oz updates
- Fixed handoff snapshot files not reaching a rebuilt cloud sandbox, which left the agent referencing files that were never downloaded. (#15505)
- Aug 27, 2026
- Date parsed from source:Aug 27, 2026
- First seen by Releasebot:Aug 27, 2026
Start from scratch, without a repo
Cursor adds a smoother Cloud Agents workflow that starts without a connected GitHub or other SCM provider, then lets users save work to a Cursor Origin repo, preview live in the browser, and publish to Vercel with a live URL.
Cloud Agents no longer require a connected GitHub or other third-party SCM provider to get started. Prompt from the get-go, then save your work to a Cursor Origin repo.
Get started
Select Start from scratch in the repo picker, then prompt the agent. In the background, Cursor creates an Origin repo for you.
Turn it into a real repo, whenever you want
Once the agent's build is to your liking, click the Create repo button to save your work in an Origin repo. Choose a custom name or pick one of the suggested ones, then set the visibility to private or internal. You'll get a fully scaffolded Origin repo, ready to share or keep building on. Navigate to the Codebase tab to find your repo and access your project.
A live preview, right in the browser
Cursor now port-forwards your cloud agent's live environment straight to your browser, so you can preview it and use tools like design mode.
Publish your work
Connect a Vercel account and hit publish to get a live URL for what you built.
A Vercel account is required to use the publish feature.
Get started today.
Original source - Aug 27, 2026
- Date parsed from source:Aug 27, 2026
- First seen by Releasebot:Aug 27, 2026
cmux by manaflow-ai
cmux-relay-v0.1.0-rc.6
cmux ships post-suspend wake-redial and read-liveness deadline fixes in cmux-relay 0.1.0-rc.6.
- Aug 26, 2026
- Date parsed from source:Aug 26, 2026
- First seen by Releasebot:Aug 27, 2026
- Aug 26, 2026
- Date parsed from source:Aug 26, 2026
- First seen by Releasebot:Aug 27, 2026