Zum Hauptinhalt springen
Antigravity

Antigravity

Vollständiges Changelog

Googles KI-gestützte Agentic IDE

Vollständiges Changelog mit allen Versionen seit dem ersten Release.

Neueste Version:v1.1.21
Releases insgesamt:99
Quelle:Web Changelog

Releases (99)

v1.1.21Neu

Antigravity CLI: Voice dictation, embedded ripgrep code search, status line cost metrics, and UTF-8 encoding fixes

26. August 2026

  • •Added /voice dictation to transcribe speech directly into the prompt with f5 and /voice toggle shortcuts, introduced the mic-serve subcommand to forward a local microphone over SSH, and improved quota messaging for concurrent stream limits.
  • •Added a cost field to the status line data model, exposing unrounded estimated session costs to track running token spend.
  • •Improved agent code search by bundling an embedded ripgrep binary instead of invoking external system binaries, increasing search performance and cross-platform consistency.
  • •Improved the always-proceed permission mode to auto-approve MCP tool calls and page reads without prompting.
  • •Improved allow-always permission suggestions for script runners (npm run, yarn, pnpm, cargo run) by scoping approvals to specific script names.
  • •Improved conversation titles by generating them automatically on creation, providing meaningful session names in the resume picker.
  • •Improved error diagnostics when an MCP server configured for Google credentials cannot locate Application Default Credentials (ADC).
  • •Fixed corrupted file edits in documents containing non-ASCII text (CJK characters, accented letters, emoji) caused by inexact offset matches producing invalid UTF-8.
  • •Fixed sessions stalling mid-response when a tool result or file diff contained invalid UTF-8.
  • •Fixed file writes being reported as failures after successfully saving to disk, preventing unnecessary agent retry loops on large files and notebooks.
  • •Fixed precedence resolution so explicitly configured skill and plugin paths take priority over automatically discovered customizations during name collisions.
v1.1.20Neu

Antigravity CLI: Skill icon branding, empty directory indexing in @-completion, and headless exit code fixes

25. August 2026

  • •Added skill icon and visual branding support across the CLI, displaying emoji icons declared under metadata.icon in SKILL.md frontmatter across the /skills catalog list view, detail inspection headers, and slash command autocompletion popups with multi-byte Unicode display width calculation.
  • •Improved @ file path autocompletion by indexing empty directories alongside files in ripgrep search results, enabling directory-only and unpopulated structures to be traversed.
  • •Improved permission management by automatically granting workspace-scoped read access under the default review mode, removing repetitive approval prompts for reading or listing workspace files while strictly maintaining confirmation for modifications and external access.
  • •Improved Git repository inspection performance by skipping recursive submodule worktree scans while continuing to track commit pointer updates.
  • •Fixed print mode (-p/--print, including --output-format json and stream-json) treating benign tool execution errors and permission denials as fatal run failures with non-zero exit codes.
  • •Fixed the CLI overwriting and discarding unparsed configuration in settings.json when encountering an unrecognized setting value or syntax error on startup.
  • •Fixed /skills, /plugins, /agents, and /hooks commands reporting that no customizations were found when invoked without an explicit agent configuration by mounting built-in customizations.
  • •Fixed CJK draft jitter in the main prompt by normalizing boundary whitespace in streaming transcript updates.
  • •Fixed the conversation spinner animation loop continuing to tick in the background after turn completion, eliminating idle CPU wakeups.
v2.10.0Neu

Embedded Terminals, Git Version Control, and Audio Attachments

24. August 2026

  • •Added an embedded terminal directly in the sidebar (Ctrl/Cmd + `), allowing you to run build commands, test suites, and scripts without switching applications.
  • •Introduced Git version control in the Review pane to inspect working tree diffs, stage and unstage files, and commit changes directly within the app.
  • •Improved Git responsiveness and reduced background CPU usage on large repositories and projects with submodules.
  • •Added support for attaching audio files (MP3, WAV, M4A, AAC, OGG, FLAC, OPUS) up to 20MB via file picker or drag-and-drop.
  • •You can now drag to select regions and leave comments directly on any image in the file viewer, with smooth scrolling for tall images.
  • •Model Context Protocol (MCP) and custom tool execution steps now display structured headers, expandable argument blocks, and formatted result previews.
  • •Redesigned pending review comments in the chat input with unified attachment badges, file-grouped hover previews, and quick modal access for queued messages.
  • •Improved sidebar smoothness and message streaming responsiveness during active generation.
  • •URL artifact cards now open supported links (such as local dev servers and Google Docs, Sheets, Slides, and Drive) directly in an in-app preview pane.
  • •Improved folder picker navigation with live directory suggestions, Windows path support, and inline new folder creation.
  • •File search in the @-mention menu is now typo-tolerant, making it easier to reference files with approximate names.
  • •Markdown file previews and artifacts now cleanly format frontmatter metadata as a structured key-value summary.
  • •Double-clicking an option in single-select questions and confirmation prompts now immediately selects and submits the answer.
  • •Fixed an issue where the slash command menu and input suggestions displayed commands unsupported by the active custom agent.
  • •Fixed an issue where goal progress and summary sections were missing from the chat interface during /goal execution.
v1.1.19Neu

Antigravity CLI: Dynamic remote control port allocation, banner logo toggles, and rendering flags

22. August 2026

  • •Added the AGY_CLI_HIDE_LOGO environment variable for narrow terminals, screen readers, and recordings to suppress logo art while retaining version and account details.
  • •Added AGY_CLI_DISABLE_ESCAPE_SEQUENCE_OPTIMIZATIONS to bypass the renderer's dirty-rectangle and diffing optimizations.
  • •Fixed --remote-control failing when a default port was occupied by automatically binding to an available port provided by the operating system.
v1.1.18Neu

Antigravity CLI: Project name flags, customizable picker keybindings, typo-tolerant @-completion, and audio attachments

22. August 2026

  • •Added support for passing a project name to --project, which previously accepted only a project ID and failed on anything else.
  • •Added item.rename and item.delete keybindings for the conversation picker's rename and delete actions, allowing F2 and F4 to be rebound in keybindings.json on keyboards without function keys.
  • •Improved @ file path completion with a typo-tolerant fallback, so a query with a transposed or mistyped character still finds the file without outranking exact matches.
  • •Improved audio attachments by recognizing standard formats Gemini models accept, including .wav, .mp3, .m4a, .aac, .flac, and .opus.
  • •Improved keystroke responsiveness on Windows by discarding key-release events reported by the console, cutting rendering work per keystroke.
  • •Improved the artifact viewer footer by placing the outline shortcut next to scroll and page hints for clearer document navigation.
  • •Fixed print mode (-p) exiting successfully with an empty response when the agent state stream dropped mid-run; it now surfaces the stream error and exits non-zero.
  • •Fixed a valueless prompt flag swallowing the next flag as its prompt (such as --print --sandbox "do the task"), turning both that and stray trailing arguments into explicit errors.
  • •Fixed expanded /btw cards pushing footer hints and prompt off-screen on long answers by clamping output to terminal height with scroll support.
  • •Fixed /resume opening below the fold instead of the active workspace when running the CLI from a subdirectory.
  • •Fixed text losing styling after file links where the link's styling reset also cleared surrounding prose formatting.
  • •Fixed stray characters appearing in the prompt every few seconds by restricting periodic input-mode re-arming to terminal multiplexers.
v0.1.14Neu

Antigravity IDE: Workspace path resolution against current working directory, compaction event deduplication, and Vertex auth overrides

21. August 2026

  • •Vertex Endpoint Authentication Overrides: Enhanced VertexEndpoint to support custom authentication headers and token overrides when routing inference requests through API gateways or proxy servers.
  • •Workspace Path Resolution: Fixed relative workspace paths in LocalAgentConfig(workspaces=[...]) incorrectly resolving against app_data_dir instead of the process working directory.
  • •Compaction Hook Deduplication: Fixed on_compaction hook dispatching multiple times per compaction event and resolved over-counting in compaction_indices by adding terminal state deduplication.
v2.9.1Neu

Remote Control, faster project switching, and more

20. August 2026

  • •Introduced Antigravity Remote Control, allowing you to securely connect to and drive agent sessions on your local device from any browser, with proactive push notifications when agents complete tasks or require input.
  • •Improved performance across the app: faster startup and project switching, faster loading of conversations with long histories, less lag during active sessions, and smoother opening and collapsing of the sidebar and side panels.
  • •Added the ability to collapse top-level sidebar sections. The ungrouped list now shows up to 30 conversations, and conversation actions appear instantly on hover.
  • •Notification toasts now appear in the bottom-right corner and collapse into a compact stack that expands on hover, so they no longer cover the screen.
  • •Hovering over a file's comment count pill now previews all comments on that file alongside their line numbers, and clicking a comment opens the referenced file, diff, or artifact with that comment focused.
  • •Refreshed the Customizations settings screen with collapsible sections and item counts for skills, rules, plugins, and custom agents, consistent back navigation across sub-pages, and clearer card layouts and status indicators for MCP server configurations.
  • •Clicking Retry on an agent error card now resumes the task directly, without adding an artificial "Continue" message to your conversation.
  • •Custom agents can now reuse the skills, rules, and subagents you already have, using a single inheritCustomizations setting in the agent's frontmatter.
  • •Added support for uploading and attaching WebP images in chat.
  • •Added syntax highlighting for PowerShell (.ps1, .psm1, .psd1, and pwsh/ps1 code fences) and Zsh (.zsh, zsh) in chat and the file viewer.
  • •Refreshed the Google Docs, Sheets, Slides, and Drive icons to the 2026 Workspace logos, and URL artifact cards for those products now show the product icon instead of a generic link glyph.
  • •Fixed an issue where Find (Ctrl+F / Cmd+F) in the Review pane missed matches that were outside the visible area or hidden inside collapsed files and folded line blocks. Search now scans all open files and automatically expands and scrolls to each match.
  • •Fixed an error that could stop you from rewinding a long conversation.
  • •Fixed an issue where a conversation could get stuck showing as still running and never return to idle, most often after pressing Stop.
  • •Large binary attachments returned by connected MCP tools — images, PDFs, and audio — are now saved to a file and referenced in the conversation instead of being pasted inline, where they could previously overflow the conversation and break the turn.
v1.1.17Neu

Antigravity CLI: Consolidated agent execution harness and media attachment fixes

20. August 2026

  • •Improved the agent execution harness by consolidating onto a single execution path, giving more consistent tool, hook, and prompt behavior.
  • •Fixed /teamwork-preview and some other slash commands disappearing for some users.
  • •Fixed Enter not opening an active background task or subagent while the prompt was in Vim insert mode.
  • •Fixed attaching Ogg audio and video files such as .ogg, .opus and .ogv, which the model rejected because they were sent as the generic application/ogg.
v1.1.16Neu

Antigravity CLI: CLI mcp management commands, debounced @ path completion, and MCP resource handling fixes

20. August 2026

  • •Added mcp subcommands (add, remove, list, enable, disable) for managing MCP servers in your user-level mcp_config.json without hand-editing it, covering both stdio and HTTP servers through --type, --env and --header.
  • •Improved @ file path completion in large workspaces by running the lookup through the bundled ripgrep and debouncing keystrokes, and by ranking a file whose name matches your query above a directory or generated artifact that merely contains it.
  • •Improved /effort so it adjusts reasoning effort for Gemini 3.6 Flash and Gemini 3.7 Flash when you sign in with a Gemini API key, a route that previously reported those models as not adjustable even though the same models were adjustable on every other sign-in path.
  • •Fixed links printing as raw escape sequences on terminals that do not implement OSC 8 hyperlinks, such as Terminal.app, and extended the same detection to command output and alert bodies, which still emitted hyperlinks unconditionally.
  • •Fixed a = character accumulating in the prompt every couple of seconds on terminals that do not implement the Kitty keyboard protocol, where the renderer's periodic re-arming of that protocol was printed as literal text instead of being interpreted.
  • •Fixed @ file path completion showing files that were in your .antigravityignore.
  • •Fixed the artifact list showing a percent-escaped filename such as quarterly%20plan.md for artifacts whose name contains a space, so the list row, the inline preview and the detail header all spell the name the same way.
  • •Fixed the prompt editor's cursor becoming misaligned when text wrapped.
  • •Fixed the /mcp panel dropping enabledTools, timeoutSeconds, url and tools.eager from mcp_config.json when you toggled a server on or off; it now preserves any field it does not recognize, so configuration written by a newer client survives an edit.
  • •Fixed read_resource discarding non-image binary content returned by an MCP server, which now offloads every blob to disk and inlines only small text and image resources, so large PDFs, audio and other binary resources are usable instead of silently dropped.
  • •Fixed Workforce Identity Federation sign-in signing you out roughly every hour, because the token refresh went to the standard sign-in endpoint rather than the federated one.
  • •Fixed skills that ship with the CLI being reported as not built-in, so the builtin flag in the /skills listing under --output-format json is accurate and the /skills panel groups them correctly.
  • •Fixed the reasoning-effort description under the timeline gauge in the /effort and /model pickers overflowing on narrow terminals, where it is now omitted so the gauge stays readable.
  • •Fixed /btw failing with a planner configuration error in conversations driven by a custom or SDK-defined agent, so side questions work regardless of which agent you are running.
  • •Fixed a failed /btw side question always reporting model returned an empty response, which was a hardcoded guess rather than the real failure; the card now shows the side question's own error when there is one.
v1.1.15

Antigravity CLI: Stream-JSON input mode, markdown agent explicit rules, and startup credential fixes

19. August 2026

  • •Added --input-format stream-json to print mode, which reads newline-delimited JSON prompts from stdin and runs one turn per message in a single conversation, so a driver can keep a session open.
  • •Added a rules: key to markdown agent frontmatter, so an agent can name rule files directly instead of inheriting your whole rule tree; a rule named this way always applies.
  • •Added plugin support for a top-level rules.json, so a plugin can declare the rule files it ships the same way it already declares its skills.
  • •Added a more specific hint on the spinner line, which now names the tool call the agent is currently running.
  • •Fixed the CLI aborting at startup on hosts that preload their own allocator through LD_PRELOAD, such as Cloud TPU VM images, where it died before the interface appeared.
  • •Fixed personal accounts hitting a resource-exhausted error at startup when saved credentials were restored from the system keyring, and cut the repeated account checks that followed.
  • •Fixed your billing project, and the license tier it selects, being lost when the CLI restarted and restored saved credentials from the system keyring.
  • •Fixed a spurious "Out of credits" message for enterprise users signed in through Application Default Credentials.
  • •Fixed the /model picker cutting off models, and its own header, on terminals too short to show every model; the list now scrolls.
  • •Fixed the /model picker misaligning its columns for model names that are not plain ASCII, by measuring column widths in display cells rather than bytes.
  • •Fixed the spinner line flickering to a generic label between quick tool calls.
  • •Fixed streamed text corrupting non-ASCII characters into replacement characters, in both the interactive display and --output-format stream-json text deltas.
  • •Fixed artifacts written into a subdirectory of a conversation's artifact directory, such as scratch/, never appearing in the artifact list.
  • •Fixed reading a .wav file failing as an unsupported media type, by normalizing non-canonical MIME types such as audio/wave before they reach the model.
v1.1.14

Antigravity CLI: Enterprise sign-in fast path, MCP OAuth metadata documents, /context scrolling, and markdown agent customizations

18. August 2026

  • •Added a faster return path for enterprise sign-in, offering returning users a single Continue with ... option naming the authentication method they last used.
  • •Added support for OAuth client ID metadata documents when connecting to an MCP server, allowing servers implementing this specification to connect without requiring a manually supplied client ID or dynamic client registration.
  • •Improved the /context panel, making it scrollable and dynamically sized to the terminal, with /usage sized the same way.
  • •Improved markdown-defined agents with a single inheritCustomizations switch that controls whether an agent adopts skills, rules, plugins, subagents, and MCP servers, replacing inconsistent per-kind defaults.
  • •Improved workspace boundary settings: accessing paths outside the workspace now grants read access only, with writes requiring authorization based on the active execution mode.
  • •Fixed the CLI exiting silently when its language server failed at startup or during execution; failures are now logged to the terminal and exit with a non-zero status.
  • •Fixed sign-in URLs and hyperlinks rendering as garbled escape sequences on terminals that do not support OSC 8 hyperlinks (such as GNU screen), detecting unsupported terminals and rendering plain text instead.
  • •Fixed Enter keystrokes being swallowed while the path-suggestion popup was loading, which previously sent the prompt and popped results open above it.
  • •Fixed the artifact list marking previously unreviewed, commented, or rejected artifacts as approved when reopening the list.
  • •Fixed malformed MCP server configurations causing all other servers to fail; invalid entries are now logged and skipped while remaining servers continue loading.
  • •Fixed built-in skills, rules, and plugins being stripped when an agent configured inherit_user: false, ensuring built-in capabilities remain available while personal customizations are excluded.
  • •Fixed conversation branching and rollbacks failing with too many SQL variables on long trajectories.
  • •Fixed conversation titles unexpectedly jumping to older messages after trajectory truncation by protecting the title checkpoint.
  • •Fixed multi-line pastes submitting individual prompts per line on terminals that do not support bracketed paste.
v0.1.13

Antigravity IDE: Pre-tool argument modification, synchronous lifecycle hooks, RunCommandConfig timeouts, and step correlation IDs

18. August 2026

  • •Pre-Tool Hook Argument Modification: Pre-tool lifecycle hooks can now sanitize, transform, or override tool input arguments before tool execution begins.
  • •Synchronous Hook Function Support: Lifecycle hook decorators now accept standard synchronous functions alongside asynchronous coroutines without raising runtime await errors.
  • •Structured Command Execution Configuration: Command execution settings are now consolidated under RunCommandConfig, introducing configurable timeouts that default to 10 minutes (600 seconds) alongside daemon execution controls.
  • •Tool Lifecycle Step Correlation: ToolResult and ToolExecutionError event payloads now include step_id, enabling end-to-end tracking and correlation of tool invocations across trajectory steps.
  • •VS Code Debugging Configuration: Updated setup_vscode_debugging.sh to target the canonical getting_started/hello_world starter example and explicitly configure Gemini Developer API defaults.
  • •Synchronous Hook Decorator Execution: Fixed a runtime TypeError when decorating synchronous functions with @pre_turn, @post_tool_call, and other lifecycle hooks by verifying awaitability before awaiting hook responses.
  • •LiteRT Early Client Disconnects: Suppressed unhandled ConnectionResetError, ConnectionAbortedError, and BrokenPipeError exceptions when clients disconnect early from local LiteRT server connections.
v1.1.13

Antigravity CLI: GEMINI_API_KEY authentication, custom agent task management, embedded ripgrep caching, and conversation database fixes

14. August 2026

  • •Added support for GEMINI_API_KEY, allowing the CLI to run directly against the Gemini API without signing in. Set modelProvider: "gemini" in settings.json, export GEMINI_API_KEY, and optionally configure GOOGLE_GEMINI_BASE_URL for custom endpoints. The banner and /help show Gemini API key as the credential, and /logout explains that it originates from the environment.
  • •Improved /codesearch resilience by falling back to a local search when the ripgrep binary cannot be executed, such as when blocked by endpoint security software.
  • •Improved artifact list scrolling performance by caching each file's preview line count instead of re-reading every listed file from disk on every keypress.
  • •Improved custom agent support by making manage_task available to declarative agents that have not opted out of fundamental components, enabling custom agents to list and kill background tasks, with clearer step descriptions like Killed task X and Checked task X.
  • •Improved search_web step visualization by displaying the query while the search is in progress rather than only after completion.
  • •Improved agent behavior after backgrounding tasks by removing anti-polling reminder text from manage_task and manage_inbox tool responses to prevent unintended polling loops.
  • •Improved the /resume picker by clearly relabeling its second tab.
  • •Improved deleting conversations in the /resume picker by remapping the shortcut from ctrl+delete to f4 for reliable cross-platform terminal compatibility on macOS.
  • •Improved the load time when opening large conversations by fetching step headers in batches without reading each step's full payload.
  • •Improved embedded ripgrep reliability and security by extracting binaries into the user cache directory instead of /tmp, adding SHA-256 integrity verification, and using atomic renaming to prevent concurrent execution races.
  • •Fixed distorted startup banner rendering in Cloud Shell by automatically enabling a conservative rendering mode for web terminals.
  • •Fixed direct Gemini API mode getting stuck on the sign-in screen instead of launching into the main UI, and resolved failures with custom GOOGLE_GEMINI_BASE_URL endpoints by omitting unsupported session fields.
  • •Fixed conversations disappearing from the /resume picker after switching tabs, and preserved the tab hint in the shortcut bar even when a tab is empty.
  • •Fixed trajectory truncation over-pruning conversation history by charging the size budget solely against reclaimable steps rather than protected checkpoints or already-cleared residues.
  • •Fixed unbounded growth of the SQLite conversation database during background task and subagent wakeups caused by duplicate grants and settings being appended to persisted rows.
v2.8.1

Chat Responsiveness Improvements

13. August 2026

  • •Improved responsiveness and reduced hanging issues when sending messages to the agent.
v2.5.5

Antigravity SDK: Windows Media Attachments and Chat Responsiveness Improvements

13. August 2026

  • •Fixed an issue where the agent failed when attaching images or media files on Windows machines.
  • •Improved responsiveness and reduced hanging issues when sending messages to the agent.
v0.1.12

Antigravity IDE: Fix proto deserialization regression for ActionGenerateImage output paths

13. August 2026

  • •ActionGenerateImage Proto Skew: Resolved client-side deserialization errors where the bundled binary emitted ActionGenerateImage.output_path while the shipped Python proto lacked the corresponding field, ensuring robust protobuf parsing across image generation steps.
v2.8.0

Persistent Sidebar Folders, Crash Fixes, and UI Enhancements

12. August 2026

  • •The sidebar now remembers which folders you have collapsed or expanded, including when you use the 'Collapse All' action, keeping your Projects list organized across sessions.
  • •Fixed an issue that prevented submitting or dismissing the feedback form on the welcome screen.
  • •Fixed frequent application crashes when opening conversations that contain large command outputs. The file viewer now limits the size of files it will display, and large changes in the history are limited to ensure stability.
  • •Fixed a styling issue where inline code in chat messages and the file viewer did not display in the correct colors.
  • •Fixed a layout bug that caused multi-line commands to display without line breaks in the activity history.
  • •Fixed an issue where the application header and top menu bar failed to display on Linux.
  • •Fixed an issue where empty section headers (such as "Terminal") appeared in Settings when all contained settings were disabled.
v2.5.2

Antigravity SDK: Queued Messages, File Attachments, and Performance Improvements

12. August 2026

  • •Improved the ask questions panel: clearer option selection, and continue now only activates once you've answered the current question.
  • •Added support for queued messages, including settings to configure execution behavior, a "Send Now" option, and an improved input card layout.
  • •Added support for attaching and rendering plain text (.txt), pdf (.pdf) and markdown (.md) files in conversations.
  • •Added a new built-in Antigravity Guide skill to provide helpful guidance when users ask about Antigravity.
  • •Improved the model selector dropdown to choose thinking levels per-model (Low, Medium, High).
  • •Updated icons for built-in slash commands and redesigned mentions/slash commands menu alignment in the chat input box.
  • •Opening long conversations is now dramatically faster.
  • •Reduced typing lag while the agent streams responses.
v2.7.1

Reliability, Performance Gains, and User Experience Polish

11. August 2026

  • •Improved performance of the conversation sidebar, especially when managing a large number of projects and conversations.
  • •Improved the performance of the conversation sidebar, reducing lag when switching between conversations.
  • •Added the ability to delete projects from the Project Settings page. Deleting a project permanently removes it along with all associated active and archived conversations.
  • •Added side-by-side previews for image files (including SVGs) when comparing file changes.
  • •Added a 'Last Prompt' sorting option to the display options menu, allowing you to sort your conversation history by the time of your last message.
  • •Improved the slash-command menu with smarter search ranking and highlighting, making it easier to find and select commands as you type.
  • •Adjusted the appearance of the thumbs-up and thumbs-down feedback buttons on messages so they no longer look disabled.
  • •The conversation hover card now displays the project name next to a folder icon, showing where the conversation is saved.
  • •You can now use the keyboard to navigate to action buttons and menus that appear when hovering over list items (such as pinning or deleting conversations).
  • •Improved sidebar organization with new conversation grouping options. You can now group conversations by project or status while keeping standalone conversations separate, or view all conversations in a single combined list.
  • •Improved the design and organization of the display options menu in the projects sidebar and history pages. Additionally, menu checkboxes have been updated to display clearer selection states.
  • •Ensured that local web servers started by the agent are reachable from your host machine.
  • •Fixed a visual issue where submenu items (such as the 'Copy' options) were vertically misaligned when opened.
  • •Fixed an issue where the chat input box lost focus after switching conversations using keyboard shortcuts.
  • •Fixed an issue where scrolling inside the file path dropdown would also scroll the main file path bar in the background.
v1.1.12

Antigravity CLI: Artifact heading outline, print-mode commands, artifact alerts/carousels, improved terminal hyperlinks, and extensive CLI fixes

11. August 2026

  • •Added a heading outline to the artifact viewer, opened with t, so a long markdown document's structure is visible at a glance and you can jump straight to a section instead of scrolling.
  • •Added non-interactive answers for more read-only slash commands in print mode, so -p "/permissions", /hooks, /help, /changelog and /config each emit one tab-separated record per line — or a structured payload under --output-format json and stream-json — without starting an agent turn, spending quota or leaving a conversation behind, with /help listing exactly the commands print mode answers and /changelog printing the release notes.
  • •Added machine-readable output to the models and agents subcommands through an --output-format flag accepting json and stream-json, and moved their error messages and progress spinner off stdout so captured output contains only the list.
  • •Added a disable-slash-command: true flag for a skill's SKILL.md frontmatter, which hides that skill from the / menu and from /name resolution while leaving it discoverable and invocable by the model, so a large skill library no longer floods the command menu.
  • •Added rendering for alert callouts ([!NOTE], [!TIP], [!IMPORTANT], [!WARNING], [!CAUTION]) and for carousel blocks in markdown artifacts, which previously showed as raw markup.
  • •Improved terminal hyperlink support, enabling clickable links over SSH and in hyperlink-capable tmux, keeping word-wrapped URLs clickable past their first line, linking URLs that the markdown renderer had left as plain text including ones inside code blocks, no longer highlighting unrelated links together on hover, and fixing horizontal scrolling and text selection on lines that contain a link.
  • •Improved tool call headers so every native and MCP tool call carries a short summary of what it is doing instead of a bare name or a raw argument blob, with the same summary naming background tasks and subagents in the activity list and the matching action phrase describing the request in permission prompts.
  • •Improved headroom for large toolsets by raising the per-session limit on tool declarations, so heavy MCP, plugin and skill setups stop being rejected for having too many tools.
  • •Improved sign-in with Application Default Credentials by resolving the quota project automatically, so service-account logins work without extra configuration.
  • •Improved repository detection so a Git repository nested inside a larger multi-repository checkout resolves to the intended root, with submodules and worktrees covered.
  • •Improved headless -p runs so the agent settles a choice itself where it would otherwise ask, instead of stalling on a question nobody is there to answer.
  • •Improved the schedule tool by dropping the redundant Timer Cancelled notification when a one-shot timer with an early-termination condition is cancelled, and by naming both the condition and the exact sender whose message satisfied it in the step result.
  • •Improved file paths in tool step lines by collapsing your home directory to ~ instead of truncating the front of the path, so the interesting end of a long path stays on screen.
  • •Fixed --mode being ignored in headless -p runs, where a valid value such as accept-edits or plan was never applied and an unrecognized value produced no warning at all.
  • •Fixed startup diagnostics being swallowed into the log file instead of reaching the terminal, including crash notices, the --conversation not-found warning, the --print and --prompt-interactive conflict, conversation load errors, and the --mode and --agent warnings.
v0.1.11

Antigravity IDE: Default Gemini 3.7 Flash upgrade, session budget limits, Vertex AI Express Mode, and autonomous behavior modes

11. August 2026

  • •Default Model Upgrade to Gemini 3.7 Flash: Upgraded the default inference model to gemini-3.7-flash.
  • •Session Budget Enforcement & Stop Reasons: Added BudgetConfig to define session-level usage limits (total tokens, turns, and cost) and StopReason enum (BUDGET_EXCEEDED, TURN_LIMIT, USER_CANCELLED, etc.) to inspect turn termination causes.
  • •Vertex AI Express Mode Support: Added native support for Express Mode authentication via VertexEndpoint(api_key=...) and LocalAgentConfig(vertex=true, api_key=...), simplifying headless and non-GCP deployments.
  • •Autonomous Agent Behavior Mode: Control the agent's behavior with AgentBehavior. By default the SDK now has an AgentBehavior.AUTONOMOUS mode (previously AgentBehavior.INTERACTIVE) to streamline scripting, background, and headless interaction modes. Override by setting CapabilitiesConfig(agent_behavior=AgentBehavior.INTERACTIVE).
  • •Multi-Interface Hook Registration: Enabled single-instance registration across multiple hook interfaces (PreToolHook, PostToolHook, PreTurnHook), allowing for cross-functional instrumentation without duplicate invocations.
  • •PreToolArgs Meta Exposed trajectory_id and step_index on tool hook payloads for chat thread context tracing.
  • •ToolRunner String Annotation Coercion: When using from __future__ import annotations, tool argument coercion failed on stringified types; resolved by resolving type annotations via typing.get_type_hints before type adaptation.
  • •MCP Server Example Port Binding: Ephemeral port race conditions in test and example server startup were resolved by binding directly to port 0.
  • •Subagent Deadlock Prevention: Handled subagent fatal errors in the localagent executor to prevent deadlocks when subagents terminate abnormally.
  • •Empty Input Validation: Added input validation to prevent SDK unresponsiveness on empty or whitespace-only prompts.
v2.6.0

Faster long conversations, more reliable hooks and subagents

7. August 2026

  • •Improved the ask questions panel: clearer option selection, and continue now only activates once you've answered the current question.
  • •Hover details on the small icon chips in chat now appear almost immediately instead of after a noticeable pause.
  • •Images and videos in chat now open in a preview window when clicked, with the option to open them in a tab instead.
  • •Hook configurations that could never run are now rejected at load time with a clear error instead of being silently ignored.
  • •Projects in the command palette now use the same folder icon as the sidebar, instead of a mismatched briefcase.
  • •Opening long conversations is now dramatically faster.
  • •You can now copy an image to your clipboard from an artifact or the file viewer, using the menu or right-click.
  • •The conversation list now loads far faster for users with a large conversation history.
  • •Settings descriptions now use consistent wording throughout, and two typos in the shortcut and host address settings were fixed.
  • •Slash commands now work when a newline or tab separates the command from its arguments.
  • •Conversations started by background automations no longer disappear from the sidebar list; they now appear under the scheduled-tasks filter.
  • •Steps waiting on your input now show the icon for the tool involved, including in the permission prompt, instead of a generic bell.
  • •Fixed the comment popup following your cursor while dragging, and whole-file comments landing on the first line with no highlight.
  • •Agents no longer hang forever when a stop hook keeps blocking the end of a turn; the turn now finishes after repeated blocks.
  • •Conversations that had already finished no longer stay stuck showing as running after you quit and reopen the app.
v1.1.11

Antigravity CLI: Vim modal editing mode, Vim-aware submission, print-mode slash-command enhancements, and various CLI and permission fixes

7. August 2026

  • •Added a Vim editing mode, off by default and switched on from /settings under Editor Mode, bringing modal editing to the prompt with Normal, Insert, Visual and Visual Line modes, a mode badge in the status line and a Vim tab in /help, and covering the h/j/k/l, w/b/e and W/B/E motions, the 0/^/$/gg/G boundaries, the f/F/t/T character searches with their ; and , repeats, the i/a/I/A insert entries, the d/c/y operators alongside x, D and C filling the unnamed register for p, and the iw/aw/iW/aW/ip/ap text objects.
  • •Added Vim-aware submission so a prompt can be sent without leaving modal editing, through an Editor Mode > Insert First toggle that opens the prompt in Insert mode where a bare Enter submits and a modified key such as shift+enter or ctrl+j inserts a newline, ctrl+s and ctrl+enter keys that submit from Normal mode, an Enter binding available as vim.insert.submit, and the full set of vim.* scopes in keybindings.json for remapping any of it.
  • •Added Vim editing to the comment editors in the diff and artifact-detail views, whose footer hints follow the active mode rather than showing one fixed set of bindings, and made a collapsed block behave as a single unit under Vim motions.
  • •Added non-interactive answers for the read-only slash commands in print mode, so -p "/usage", /quota, /credits, /model, /effort and /skills emit one tab-separated record per line — or a structured payload under --output-format json and stream-json — without starting an agent turn, spending quota, or leaving a conversation behind.
  • •Added an explicit refusal for the remaining interactive-only slash commands in print mode, which previously fell through as literal prompt text and let the model answer as though the command had run, so -p "/clear" reported the context cleared while nothing was cleared; each now fails with the flag or subcommand that replaces it.
  • •Improved plugin enable and disable so config.json is the only place enablement lives, seeded once from each plugin's manifest, which stops a plugin that later ships "disabled": true from switching itself off under someone who was already running it and stops a shipped-default change from moving every user on the next release.
  • •Improved the artifact detail view by wrapping long lines in code files instead of clipping them at the right edge.
  • •Improved model retries by honoring the server-supplied retry delay instead of the client's own backoff, so a retry after a rate limit or an overload waits exactly as long as the server asks.
  • •Improved model-loading errors so a permission failure such as a missing license or a missing IAM role is shown as itself, with a troubleshooting link, instead of a generic failure during the run.
  • •Fixed an allowlist entry that tokenizes to zero command words — command(time), a comment-only entry, or an empty compound such as () — matching every command and silently auto-approving anything the agent ran; such an entry now matches nothing.
  • •Fixed commands being auto-approved while the session was in request-review or strict permission mode.
  • •Fixed admin controls being skipped for MCP servers at startup, where a fetch made before authentication cached "admin controls not applicable" and allowed every server for the next five minutes, and fixed the built-in Chrome DevTools MCP server being blocked outright by admin controls.
  • •Fixed MCP progress callbacks being dropped and the task log file not being initialized, so long-running MCP tool calls report progress again.
  • •Fixed a slash command receiving the collapsed [Pasted text #N] placeholder instead of the pasted content when its argument was pasted into the prompt.
  • •Fixed the prompt saving the typed prefix rather than the completed command name to history when an autocompleted slash command was submitted, so up-arrow recall replays what actually ran, and made an alias require an exact match before auto-executing so a partial alias fills the prompt instead of running a command you did not finish typing.
v0.1.10

Antigravity IDE: Prioritized inference tiers, real-time token usage streaming, and context-aware hook decorators

4. August 2026

  • •Gemini Prioritized Inference Service Tier: Configure agents to utilize Gemini Prioritized Inference service tiers for high-priority model execution with automated graceful fallback.
  • •Tool Call ID Correlation in Lifecycle Hooks: Inspect call_id attributes on tool executions, errors, and hooks to correlate multi-step tool invocations across lifecycle callbacks.
  • •Context-Aware Hook Decorators: Decorate hook handlers (@hooks.pre_turn, @hooks.post_tool_call, etc.) that optionally accept HookContext as a parameter to maintain state and share data across lifecycle callbacks.
  • •ActionCompaction Event Emission & Hook: Track context window compaction notifications over WebSockets and intercept them using @hooks.on_compaction.
  • •Standardized System Instructions Strategy: Plain string instructions default to appending to built-in instructions. To override and completely replace built-in instructions, pass CustomSystemInstructions.
  • •Live Token Usage Reporting: Introduced real-time UsageUpdate event streaming so token usage accumulates live during agent execution rather than delaying updates until state transitions.
  • •Interactive CLI Spinner: Updated CLI interactive loop spinner to list all active tool names when running concurrent tool calls (e.g., Running tools 'tool_a', 'tool_b').
  • •Module Re-exports: Re-exported ReadUrlContentResult and SearchWebResult in connections.local for uniform tool result access.
  • •Local Gemma Model Documentation: Added guides and tutorials for running agents locally with Gemma models using LiteRT and OpenAI-compatible endpoints.
  • •LiteRT Token Output Limit: Increased max_output_tokens default in LiteRT local server configuration from 8,192 to 16,384 tokens to prevent truncation during complex reasoning and generation tasks.
  • •ActionCompaction Event Emission: Fixed issue where compaction notifications were suppressed in external SDK releases, causing @hooks.on_compaction handlers and conversation.compaction_indices tracking to fail; compaction events now emit properly over WebSockets.
  • •MCP Test Server Startup: Fixed a race condition where the HTTP port was exposed before uvicorn server startup completed, which previously caused intermittent ConnectionRefusedError failures during test initialization.
v1.1.10

Antigravity CLI: Gemini Enterprise & WIF authentication, read-only .git sandbox rules, hook ordering improvements, and subagent and CLI fixes

3. August 2026

  • •Added Business sign-in for Gemini Enterprise accounts, so you can authenticate with a Google Cloud project under Google Cloud terms, use a license seat allocated or auto-assigned from your organization's GE-Standard or GE-Plus subscription, run inference in a chosen region, and have your organization's administrator controls applied to various features.
  • •Added Workforce Identity Federation sign-in for enterprise users, available as the Use advanced SSO config option on the Google Cloud sign-in screen, so organizations that federate identity through an external provider can authenticate with their own identity provider.
  • •Added sign-in with Application Default Credentials so you can use Agent Platform.
  • •Added a non-blocking advisory banner when the same conversation is already open in another CLI instance on the same machine, pointing at /fork so two sessions no longer interleave writes into one trajectory.
  • •Improved the terminal sandbox by granting read-only rather than writable access to a Git repository's .git directory, so the agent can inspect repository metadata without being able to rewrite it from inside the sandbox.
  • •Improved hook ordering so hooks defined in hooks.json run before the built-in termination checks, which lets PostInvocation hooks observe the final invocation of a turn and lets Stop hooks run at all instead of sitting unreachable behind the built-ins.
  • •Improved the schedule tool to accept DurationSeconds and MaxIterations when a model emits them as bare JSON numbers rather than strings, accepting integral values and rejecting non-integral ones with a clear error instead of failing the call.
  • •Fixed --model and --effort being ignored in interactive sessions and in headless -p runs, where the flags were applied after model configuration had already been initialized so the run silently fell back to the persisted or default model.
  • •Fixed a bare --effort resolving against the default model instead of the model you actually have selected, which could silently move you to a different model.
  • •Fixed stopping a subagent tree stopping only the conversation it was invoked from, while every descendant subagent and the background tasks they owned kept running and the CLI still reported them as killed.
  • •Fixed a forced-continuation deadlock where a coordinator waiting on active subagents or background tasks would loop injecting empty continue steps until it hit the invocation limit, wasting tokens and blocking progress.
  • •Fixed the spacebar not toggling an option in multi-select prompts, including the ask_question dialog and the onboarding import checkbox, which left x as the only working toggle; the hint bar now advertises it.
  • •Fixed the Left and Right arrow keys being captured to navigate the input box suggestion dropdown, so you can move the cursor and edit text again while suggestions are showing.
  • •Fixed the model picker's "No models available" state rendering without its header and footer, so it now shows the standard chrome and an esc hint to go back.
  • •Fixed tools that an MCP server marks to always run in the background executing as blocking calls that stalled the turn.
v2.5.0

Enterprise sign-in: Gemini Enterprise accounts and Workforce Identity Federation

31. Juli 2026

  • •Added support for signing in with your Gemini Enterprise user account, including support for organization admin controls.
  • •Allocated and auto-assigned license seats against GE-Standard & GE-Plus organization subscription tiers.
  • •Regionalized (US, EU, Global) model inferencing.
  • •Operating under Google Cloud terms of service.
  • •Added Workforce Identity Federation sign-in for enterprise users: select the Advanced SSO login option.
  • •Added per-model reasoning effort levels (Low, Medium, High).
  • •Added enterprise admin policies for browser features, MCP allowlists, and permission thresholds.
  • •Approved permissions are now remembered for the rest of the conversation.
  • •Sidecar links now render as pills that open the sidecar in the auxiliary pane.
  • •Added "Reveal in Finder/Explorer/File Manager" to the file viewer.
  • •Added "Collapse All Folders" and "Expand All Folders" to the command palette.
  • •Added search and keyboard navigation to the project selector.
  • •Added syntax highlighting for Objective-C, Swift, and Dart.
  • •External markdown links now reveal the destination URL on hover.
  • •The environment selector now remembers your last-used worktree.
v1.1.9

Antigravity CLI: Slash-command expansion in print mode, non-blocking MCP startup, session-wide permission memory, system temp directory write grants, and hook and UI stability fixes

31. Juli 2026

  • •Added slash-command and skill expansion to print mode, so a headless run such as -p "/my-skill review this diff" now resolves and applies the skill instead of sending it as literal text, with --disable-slash-commands to opt out.
  • •Improved interactive startup so a slow or hanging MCP server no longer stalls the first agent turn, loading MCP servers in the background for the interactive session while headless and one-shot runs keep blocking so their single scripted turn still sees the full toolset.
  • •Improved permission grants so a pattern approved at a prompt is recorded for the rest of the conversation, letting later commands that match it run without prompting again.
  • •Improved the default system temporary-directory grant to cover writes as well as reads, so agents no longer trigger a permission prompt when creating or updating files there.
  • •Fixed stop hooks that always block hanging the agent forever; after a configurable number of consecutive continuations, the hook can no longer block and the turn ends normally.
  • •Fixed PostToolUse hooks firing on non-tool steps such as user input and model responses, which also caused them to ignore their configured matchers.
  • •Fixed slash commands not being recognized when separated from their arguments by a newline or tab, so a prompt starting with a command followed by a newline is now parsed as a command instead of being sent verbatim.
  • •Fixed deleting into a collapsed paste placeholder removing one character at a time, which left a visible fragment in the prompt while the full pasted content was still submitted; the block is now deleted atomically.
  • •Fixed the artifact viewer losing syntax highlighting when returning from the editor view, and returning to the wrong panel when exiting the artifact detail view.
  • •Fixed the headless stream-json init event advertising tools that are not available in your build.
  • •Fixed MCP servers forcing a full re-authentication after a dropped connection.
v2.4.3

Preview tabs, MCP timeouts, file attachments, and performance fixes

28. Juli 2026

  • •Added a keyboard shortcut (Cmd+L / Ctrl+L) to quickly quote selected text into the chat input.
  • •Added support for attaching .json and .md files to the input box with distinct visual badges.
  • •Added preview (temporary) tabs to the auxiliary pane, allowing files to open in a temporary slot that is reused until promoted to a permanent tab (e.g., via double-click).
  • •Added a timeout for MCP server connections and tool calls to prevent the agent from hanging indefinitely.
  • •Added an 'Only Unread' filter option to the Conversation History dropdown.
  • •Added support for uploading and attaching CSV (.csv) files in chat as plain text.
  • •Added a 'Download Diagnostics' command to the command palette to download a diagnostics package (containing logs and agent state) for troubleshooting.
  • •Added tooltips to file links in markdown views, showing their full path on hover.
  • •Refined command approval rules to reduce redundant permission prompts for everyday tools.
  • •Improved the speed of conversation switching by optimizing trajectory summary list selectors.
  • •Improved authentication error messages to show the specific failure reason.
  • •Draft comments are now automatically saved and preserved across sessions.
  • •Improved interface responsiveness by preventing unnecessary re-renders when toggling sidebars or auxiliary panes.
  • •Improved accessibility in the Settings modal with better heading hierarchy and keyboard focus management.
  • •Updated the default settings screen name from 'Agent' to 'General' and updated copy for inherited settings ('Use Global' to 'Inherit General').
v1.1.8

Antigravity CLI: Print mode structured output formats, custom JSON schema enforcement, copyOnSelect TUI setting, and compound-command permission improvements

28. Juli 2026

  • •Print mode (-p / --print) now supports structured, machine-readable output via the --output-format flag (text (default), json, or stream-json), so headless runs in CI, eval harnesses, and scripts can consume the CLI's output programmatically; these flags are now discoverable in --help.
  • •Added the stream-json output format: a strongly-typed NDJSON event stream that emits typed init, step_update, and terminal result events with a stable, closed-vocabulary step_type discriminator, so consumers receive progress incrementally instead of waiting for the whole run to finish.
  • •Added the --json-schema flag to enforce a custom JSON schema on the structured output, accepting either an inline schema string or a path to a schema file; for stream-json the schema applies to the final result event.
  • •Enriched the structured stream with a tool_info object for each tool call (canonical tool name, parameters, and output) and a subagent_info payload for delegated subagents (including conversation_id and log_uri) so consumers can correlate child trajectories.
  • •The JSON usage object emitted by json and stream-json now reports token accounting including cache_read_tokens, so non-interactive consumers can attribute prompt-cache hits.
  • •Added a copyOnSelect setting (default on, toggleable in /settings) that controls whether releasing a mouse text-selection auto-copies it to the system clipboard in the TUI's altscreen rendering mode; disable it to stop the automatic copy on release — useful when the auto-copy is unwanted or corrupts certain payloads.
  • •Improved compound-command permissions so an exact chained command (such as git fetch && git rebase) can be saved as an allow-always rule and no longer re-prompts on the next identical run.
v0.1.9

Antigravity IDE: Model-call retry/backoff configuration, LiteRT warm-up scaling, and tool execution error handling

27. Juli 2026

  • •Model-Call Retry & Backoff Configuration: Exposes configurable retry and backoff parameters in LocalAgentConfig for model calls.
  • •LiteRT Warm-up Timeout Scaling: Dynamically scales LiteRT engine warm-up timeouts based on context size and synchronizes warm-up cleanup.
  • •Improved Tool Registration: Enhances automatic tool registration and docstring parsing for custom Python functions.
  • •BuiltinTools Typing Exports: Exports BuiltinTools at top-level package and typing boundaries for subagent tool configuration.
  • •Connection DebugConfig: Adds base DebugConfig options for enhanced connection debugging and logging.
  • •Audio Payload Processing: Fixes processing for user audio payloads in interactive agent sessions.
  • •ToolExecutionError Handling: Introduces ToolExecutionError to SDK error types for explicit tool failure handling.
Zeige 30 von 99