feat(chat): add live voice input waveform

[waleedlatif1]

What

• replace the active microphone glyph with a compact five-bar live waveform clipped to the existing circular control
• use neutral EMCN active-state tokens, reduced-motion handling, and ref-driven SVG updates
• add an actionable desktop recovery path that opens native microphone privacy settings after access is denied

Why

Voice input should visibly confirm captured speech without expanding the composer or introducing an off-system alert color. A denied desktop permission should also have a direct recovery action.

How

• derive smoothed RMS levels from the existing PCM callback and expose them through a stable ref
• animate five SVG lines with requestAnimationFrame, avoiding React rerenders on audio frames
• gate a fixed native settings URL behind trusted-origin desktop IPC and preserve backward compatibility for older shells

Test Plan

• bun run lint:check
• bun run check:audits
• bun run apps/sim/scripts/check-block-registry.ts origin/staging
• bunx turbo run type-check
• bun run test
• bun run check:desktop-bridge
• bun run check:desktop-ipc
• visual verification in dark theme with live waveform samples

[vercel]

The latest updates on your projects. Learn more about Vercel for GitHub.

1 Skipped Deployment

Project	Deployment	Actions	Updated (UTC)
docs	Skipped		Aug 27, 2026 1:54am

[greptile-apps]

Greptile Summary

The PR adds a ref-driven live microphone waveform and a desktop recovery action for denied microphone access.

• Derives smoothed audio levels from captured PCM data and animates five SVG waveform bars without React frame-by-frame rendering.
• Adds reduced-motion behavior and updates the microphone control’s active styling and accessibility state.
• Adds an origin-gated desktop IPC operation for opening native microphone privacy settings on macOS and Windows.
• Hides the recovery action on unsupported platforms and preserves compatibility with older desktop shells.

Confidence Score: 5/5

The PR appears safe to merge.

No blocking failure remains; the Linux preload omits the unsupported capability, the renderer preserves that absence, and the recovery action is rendered only when the capability is available.

Important Files Changed

Filename	Overview
apps/desktop/src/main/ipc.ts	Adds an app-origin-gated IPC handler that opens fixed native microphone-settings URLs only on supported operating systems.
apps/desktop/src/preload/index.ts	Exposes the optional microphone-settings capability only on macOS and Windows, resolving the previously reported Linux no-op action.
apps/sim/app/workspace/[workspaceId]/home/components/user-input/user-input.tsx	Adds the recovery action only when the desktop bridge exposes the corresponding platform capability.
apps/sim/app/workspace/[workspaceId]/home/components/user-input/components/mic-button/mic-button.tsx	Replaces the active microphone glyph with an accessible, reduced-motion-aware SVG waveform animated through refs.
apps/sim/hooks/use-speech-to-text.ts	Computes and exposes five smoothed RMS audio levels through a stable ref while retaining the existing PCM capture flow.
packages/desktop-bridge/src/index.ts	Adds the microphone-settings operation as an optional bridge capability for backward compatibility.

_{Reviews (2): Last reviewed commit: "fix(desktop): hide unsupported microphon..." | Re-trigger Greptile}

[waleedlatif1]

@greptile

[cubic-dev-ai]

No issues found across 10 files

Confidence score: 5/5

• Automated review surfaced no issues in the provided summaries.
• No files require special attention.

_{Re-trigger cubic}