Skip to content
Proposals/Enable /voice dictation and pin allow-always to
proposaldeveloperP5Worth a lookGoogle Antigravity

Enable /voice dictation and pin allow-always to named package scripts

Antigravity 1.1.21 adds /voice (f5) dictation, mic-serve for SSH remotes, a status-line cost field, always-proceed coverage for MCP and page reads, and script-pinned allow-always for npm/yarn/pnpm/cargo. Adopt those so prompts, spend, and grants stay tight.

Why this loop

1.1.21 changes how you talk to the agent and how much access you grant it. /voice (or f5) transcribes speech straight into the prompt, and mic-serve forwards a local microphone to a CLI running on another machine over SSH; only one live dictation stream is allowed per account, and a second stream now reports that limit instead of an exhausted quota. The status line data model now exposes an unrounded estimated session cost so a custom status line can show running token spend. always-proceed finally auto-approves MCP tool calls and page reads, matching the no-interruption intent of that mode. Allow-always suggestions for npm run, yarn, pnpm, and cargo run now pin the specific script name, so approving npm run dev no longer grants every script in the project. Explicit skill and plugin paths now win collisions against auto-discovered customizations, and code search uses an embedded ripgrep so results match even when the machine has no system ripgrep. Non-ASCII edit corruption, UTF-8 session stalls, and false write-failure retries on large files are also fixed.

Proposed actions

  1. Upgrade Google Antigravity to 1.1.21, then press f5 or run /voice to start dictation (speech is transcribed straight into the prompt) and press f5 or run /voice again to stop. Keep only one live stream per account.
  2. On the machine that has the microphone, run Antigravity's mic-serve subcommand so that local mic is forwarded over SSH to the Antigravity CLI running on the remote machine.
  3. Update your custom status line to read the new cost field from the status line data model and display the unrounded estimated cost of the current session.
  4. Enable always-proceed permission mode in this Antigravity session so MCP tool calls and page reads auto-approve instead of prompting.
  5. When Antigravity offers allow-always for a script runner, approve only the pinned command (npm run <script>, yarn <script>, pnpm <script>, or cargo run) so the grant does not cover every script in the project. Put skill and plugin paths you care about in explicit config so they win name collisions.

Agent prompt

Paste into your agent or query via MCP (get_agent_prompt) — free, no extra AI cost

Google Antigravity / agent task

DevAgentRadar → Google Antigravity

Goal: turn this release signal into a safe pilot plan for my stack.

Context

Assistant: Google Antigravity Proposal: Enable /voice dictation and pin allow-always to named package scripts Summary: Antigravity 1.1.21 adds /voice (f5) dictation, mic-serve for SSH remotes, a status-line cost field, always-proceed coverage for MCP and page reads, and script-pinned allow-always for npm/yarn/pnpm/cargo. Adopt those so prompts, spend, and grants stay tight. Primary source: https://github.com/google-antigravity/antigravity-cli/releases/tag/1.1.21

Why it matters

1.1.21 changes how you talk to the agent and how much access you grant it. /voice (or f5) transcribes speech straight into the prompt, and mic-serve forwards a local microphone to a CLI running on another machine over SSH; only one live dictation stream is allowed per account, and a second stream now reports that limit instead of an exhausted quota. The status line data model now exposes an unrounded estimated session cost so a custom status line can show running token spend. always-proceed finally auto-approves MCP tool calls and page reads, matching the no-interruption intent of that mode. Allow-always suggestions for npm run, yarn, pnpm, and cargo run now pin the specific script name, so approving npm run dev no longer grants every script in the project. Explicit skill and plugin paths now win collisions against auto-discovered customizations, and code search uses an embedded ripgrep so results match even when the machine has no system ripgrep. Non-ASCII edit corruption, UTF-8 session stalls, and false write-failure retries on large files are also fixed.

Suggested actions

  1. Upgrade Google Antigravity to 1.1.21, then press f5 or run /voice to start dictation (speech is transcribed straight into the prompt) and press f5 or run /voice again to stop. Keep only one live stream per account.
  2. On the machine that has the microphone, run Antigravity's mic-serve subcommand so that local mic is forwarded over SSH to the Antigravity CLI running on the remote machine.
  3. Update your custom status line to read the new cost field from the status line data model and display the unrounded estimated cost of the current session.
  4. Enable always-proceed permission mode in this Antigravity session so MCP tool calls and page reads auto-approve instead of prompting.
  5. When Antigravity offers allow-always for a script runner, approve only the pinned command (npm run <script>, yarn <script>, pnpm <script>, or cargo run) so the grant does not cover every script in the project. Put skill and plugin paths you care about in explicit config so they win name collisions.

Config surfaces this release may change

  • permission rules (high confidence) — check your repo before applying
  • MCP servers — check your repo before applying

After you finish

Do not report this as applied to DevAgentRadar. You cannot write the visitor's loop.

Tell the human: open https://devagentradar.com/proposals/google-antigravity-1-1-21-enable-voice-dictation-and-pin-allow-always-to and mark Applied, Skipped, or Failed. Proposal id: 995ae23d-b3af-4baf-82d3-96628b2888d2

Please:

  1. Map the change to concrete pilot steps
  2. Flag security / permission implications
  3. Keep the pilot reversible
agentmcpmodelsecuritycliRelease source ↗

Your loop

This browser · no sign-in · not shared as “you”

After you run the prompt

Only you can mark this. Agents cannot write your loop.

Your decision stays on this device. A public tally appears after a few votes.

Originating release signal

Google Antigravity1.1.21Aug 26, 2026

1.1.21

Added /voice dictation, which transcribes speech straight into the prompt; press f5 or run /voice to start and stop it, and use the mic-serve subcommand to forward a local microphone to a CLI running on another machine over SSH. · Only one live dictation stream is allowed per account, and a second one now says so instead of reporting an exhausted quota. · Added a cost field to the status line data model, exposing the unrounded estimated cost of the current session so a custom status line can display running token spend. · +9 more changes
Verified excerpt — the source's own words
  • Added /voice dictation, which transcribes speech straight into the prompt; press f5 or run /voice to start and stop it, and use the mic-serve subcommand to forward a local microphone to a CLI running on another machine over SSH. Only one live dictation stream is allowed per account, and a second one now says so instead of reporting an exhausted quota.
  • Added a cost field to the status line data model, exposing the unrounded estimated cost of the current session so a custom status line can display running token spend.
  • Improved the agent's code search by running an embedded ripgrep binary instead of shelling out to whatever the machine provides, so searches are faster and behave identically on systems with no ripgrep installed.
  • Improved the always-proceed permission mode to auto-approve MCP tool calls and page reads as well, which kept prompting even though the mode exists to run without interruptions.
  • Improved allow-always permission suggestions for script runners such as npm run, yarn, pnpm and cargo run by pinning the specific script name, so approving npm run dev no longer grants every script in the project.

Excerpt ends here — this release continues at the source ↗.

Primary source ↗