Google Voice Access vs Superwhisper

Google Voice Access lets people control an Android device by speaking, including people who have difficulty manipulating a touchscreen. It can open apps, go back or home, show notifications and operate the screen currently in use. Users can display numbers or text labels beside screen items and speak the matching marker to select one. Voice Access also supports dictation and text editing, including replacing words or phrases. Google lists English, Spanish, German, Italian, French, Portuguese and Japanese; Portuguese and Japanese cover only the top 25 commands. Offline use requires downloading a speech-recognition language pack to the device. Speech is mainly processed on-device, but audio may be sent to Google servers depending on device settings and system language; Google says audio sent for this service is not saved. Audio and transcripts are not stored unless the user enables “Help improve Voice Access Commands.” The microphone listens while Voice Access is active. It does not identify a particular speaker, so anyone can control an awake, unlocked device while it is listening.

Superwhisper turns speech into formatted text for use across applications. Its Super Mode reads the screen and adjusts dictated text to the current app, for example treating email as a message or coding prompts technically. It supports more than 100 languages, automatic language detection, and translation between languages. On-device speech models process audio locally, work without internet, and keep audio off the device's outbound path. With cloud models, audio and text pass through the Superwhisper proxy to the provider; the product says they are not retained for training. It works wherever text can be entered and names Slack, Cursor, and Notion, with coding plugins for Claude Code, OpenCode, Codex, and Pi. A CLI can search and export transcripts, report recording statistics, and start a local MCP server. Free includes voice-to-text, meeting recording and transcription, and email support. Pro costs $8.49 per month or $84.99 per year and adds audio and video transcription, cloud and local models, custom API keys, and priority support. Pro features can be tried for 3,000 words before free-tier features remain available.