Tag Archives: private dictation

RITS AI Voice Bridge v3.0

Posted on by
Reply

AI Voice Bridge v3.0 is a free privacy-focused local utility from Reasonable IT Service that runs on your computer and adds a private voice layer to ChatGPT, Claude, DeepSeek, and Gemini in Google Chrome.

It supports local dictation, local readback, a floating Mic | Repeat | Stop control, voice command auto-send, and cleaner installer packaging for both Windows and macOS.

  • No cloud TTS
  • No API key
  • No extra account
  • No telemetry
  • No startup service
  • No platform-controlled voice mode required

Your AI writes. Your computer talks. You stay in control.

Note: Google Chrome and Python are required.

What’s New in v3.0

  • Support for ChatGPT
  • Support for Claude
  • Support for DeepSeek
  • New support for Gemini
  • Local dictation with Whisper
  • Read-aloud using your computer’s built-in system voice
  • Floating Mic | Repeat | Stop control
  • Repeat button to replay the latest AI response
  • Stop button to interrupt local readback
  • Voice command auto-send support
  • Improved ChatGPT, Claude, DeepSeek, and Gemini handling
  • Improved DeepSeek readback handling so final answers are prioritized cleanly
  • Cleaner Windows and macOS installer packs
  • New app icon and Chrome extension icon assets
  • Windows Desktop and Start Menu shortcuts
  • macOS Desktop launcher app

What It Does

AI Voice Bridge runs locally on your computer.

It adds a small floating control to supported AI chat sites in Chrome. You can dictate a prompt locally, insert it into the chat box, send it manually, or use a simple final phrase command to send it automatically.

When the AI responds, AI Voice Bridge can read the response aloud using your computer’s local system voice.

It is not trying to replace ChatGPT, Claude, DeepSeek, or Gemini.

It adds a local voice bridge on top of the normal text chat experience.

Voice Command Auto-Send

AI Voice Bridge supports a simple voice command for sending messages.

If your dictation ends with “send message” or “send a message”, the bridge removes that final command phrase and submits the remaining prompt automatically.

Example:

You say: Explain this in one paragraph send message

The AI site receives: Explain this in one paragraph

The command only triggers when the send phrase is the final phrase. Saying something like “send message to the client” will not trigger auto-send.

Controls

AI Voice Bridge keeps the controls simple.

  • Click the Mic button to start recording
  • The Mic button turns red while recording
  • Click the Mic button again to stop recording and transcribe
  • Press Enter while recording to stop and transcribe
  • Windows dictation hotkey: Ctrl + Shift + V
  • Mac dictation hotkey: Cmd + Shift + V
  • Click Repeat to replay the latest captured AI response
  • Click Stop to stop current local readback
  • Press Esc inside the AI site tab to stop local readback

Why We Built It

Native voice modes can be convenient, but they come with tradeoffs.

The biggest one is control. With native voice, your audio is routed through the platform’s voice system. That means the voice feature, transcription behavior, retention rules, usage limits, and overall experience are controlled by the platform.

For example, OpenAI’s own Voice Dictation FAQ says audio from dictation is retained for as long as the chat remains in your chat history. Its Voice Mode FAQ says audio and video clips from voice chats are retained for 30 days.

That may be fine for some people, but we wanted another option.

With AI Voice Bridge, dictation audio is sent only to the local server running on your own computer. The supported AI platform receives the final text prompt only when you choose to send it.

Voice features can also change. They can be limited, paywalled, removed, tuned differently, or made less useful over time. The voice experience you like today may not be the same one you get tomorrow.

There is also a practical difference between using a platform’s native voice mode and using the normal text version of the AI. Native voice can feel different from the regular chat experience. Sometimes it does not feel like you are using the same AI workflow.

AI Voice Bridge takes a simpler approach.

  • Use the normal AI chat in Chrome
  • Dictate locally on your own machine
  • Send only the final text prompt when you choose to send it
  • Let the AI answer in text
  • Let your own computer read the answer out loud locally

That’s the point. Less platform control. More user control.

How It Works

AI Voice Bridge runs a small local server on your computer at 127.0.0.1:5055.

The Chrome extension talks to that local server.

When you dictate, your microphone audio is sent to the local server on your own machine, transcribed locally with Whisper, and inserted into the AI chat box as text.

When the AI responds, the bridge captures the response text and reads it aloud through your computer’s local system voice.

The supported AI platform receives the same thing it normally receives: a text prompt.

Raw microphone audio is not sent to ChatGPT, Claude, DeepSeek, or Gemini by AI Voice Bridge.

Why It Matters

This is not trying to replace the AI platforms.

It gives you a local voice layer for the AI tools you already use.

That means you can keep using normal ChatGPT, Claude, DeepSeek, and Gemini in Chrome while keeping voice interaction simple, local, and under your control.

Native voice mode is convenience. AI Voice Bridge is control, privacy, and consistency.

AI Voice Bridge v3.0 is available now for both Windows and Mac.

Updated demo video: coming soon…

AI Voice Bridge is not affiliated with, endorsed by, or sponsored by OpenAI, Anthropic, DeepSeek, or Google. ChatGPT, Claude, DeepSeek, and Gemini are trademarks of their respective owners.