Key takeaways

Google announced voice mode for Antigravity on November 6, 2026.

The feature uses Gemini Live in the Antigravity App and CLI 1.5.0 and later.

Users can speak tasks, interrupt replies, and ask for status while work continues.

Some actions still need on-screen approval, including destructive actions and force-pushes.

The source gives no Morocco-specific availability commitment.

Google's Antigravity blog announced a new voice mode built on Gemini Live. The feature is available in the Antigravity App and in Antigravity CLI version 1.5.0 and later. It lets users speak tasks, steer ongoing work, ask for status, and receive spoken replies.

The post says Gemini Live hears the user's voice directly, rather than only receiving a transcript. That matters because the interaction is meant to feel more immediate. The existing conversation still carries context, while code edits, commands, and tool activity remain visible in that same conversation.

What Google announced

The default behavior is quiet operation until the system has an answer or needs input. Users can also request periodic spoken updates. In the app, Google says users can choose from nine voices and set speaking speed from 0.75x to 2x normal.

The CLI includes voice controls as well. Users can start voice mode with `/voice live` or F6. They can also configure voice, speed, and interruption behavior through commands.

Google also described remote SSH support. In that setup, a local microphone can be relayed through `agy mic-serve` and the `ANTIGRAVITY_MIC` setting. Ending voice mode preserves the conversation, and it does not stop a task that is still running.

Guardrails and approval flow

The source says ordinary permission settings still apply. Some permission requests can be answered by voice. However, destructive actions, elevated commands, and force-pushes still require on-screen approval.

Users can still deny those actions by voice. That is an important operational detail, because the voice layer does not replace the existing permission model. These product guardrails do not amount to a guarantee that every action is safe.

Why this matters for developers

For developers who already use Antigravity, voice mode may reduce friction during long sessions. It could help when a user wants to check progress, redirect a task, or add a correction without switching away from the workflow. It may also be useful when hands-free interaction is preferred.

The practical value depends on the task and the environment. A spoken interface can speed up simple coordination. It can also create new attention demands if users rely on frequent updates or interruption-heavy workflows. The guardrails suggest that sensitive actions still need careful review.

What this means for Morocco

The source does not give a Morocco-specific availability commitment. So the safest reading is limited to the product itself and its stated controls. For Moroccan readers, the immediate relevance is for developers, technical teams, and CLI users who already work with Antigravity and want to understand the new interaction model.

If the feature becomes available to a user in Morocco, the main benefit would be workflow control through speech. That could matter in situations where a developer wants to keep a task running while asking for status or making small corrections. It could also help teams that prefer a conversational interface for monitoring work.

The constraints are just as important. The source says destructive actions, elevated commands, and force-pushes still need on-screen approval. That means voice mode is not a full replacement for visual review. It is an added control layer on top of the existing permission system.

For Moroccan readers evaluating adoption, the key question is whether voice interaction fits their workflow. The feature is designed to preserve conversation context while keeping code and tool activity visible. That combination may suit users who want faster coordination without losing traceability. It may be less useful for users who prefer fully silent, screen-first work.

Bottom line

Google is extending Antigravity with a voice-first control layer powered by Gemini Live. The feature adds convenience, but it keeps approval checks in place for sensitive actions. For Moroccan readers, the news is most relevant as a workflow update for developers and CLI users, not as a confirmed local rollout.