Text-to-Mic: Free AI Text-to-speech-to-microphone TTS & STTTS App

Text-to-Mic is an open-source, free text-to-speech and speech-to-text-to-speech (TTS and STTTS) to-microphone tool that turns typed text into speech audio with AI, then plays that audio to your speakers, headset, or microphone feed.

Here is a video example of how it looks when running on Windows: watch the demo (MP4) or watch it on YouTube.

This is perfect to enable you to speak in online video meetings using text-to-speech AI. It can also manipulate text with AI in real time, which has lots of practical uses, such as tidying up speech or live translation. (See the download links below.)

Text-to-Mic uses the OpenAI text-to-speech engine, which surpasses the standard text-to-speech tools available on Windows and Mac. This app is available to use for free.

Seamless text-to-speech-to-microphone (or speakers) conversion. Uses OpenAI's API to convert text into natural-sounding speech in real time.
Multiple voices. Choose from a variety of OpenAI voices to find the tone that best suits your presentation or meeting style. Supported voices: Alloy, Echo, Fable, Onyx, Nova and Shimmer (listen to samples).
Customisable tones. Take control of not just the accent of the voice, but the tone and the way it speaks too. Text-to-Mic comes pre-loaded with some tones you might enjoy, and you can add your own.
Dual output capability. Outputs audio simultaneously to both headphones and a virtual microphone, so you can monitor and share your presentation at the same time.
STTTS, speech-to-text-to-speech. Record your voice, even if you are struggling to speak; it is saved as text, which you can then immediately play back over the selected audio feeds.
Hotkeys for quick access. Trigger speech recording, conversion and playback with hotkeys (like ctrl+shift+0) to make using Text-to-Mic feel quick and natural.
Automatic AI copyediting. Tidy up, manipulate or translate what you have typed or recorded into another language, or transform the input text in some other way, speeding up the communications process.

Watch the demo video linked above to see Text-to-Mic in action.

If you like this tool, we also have a free speech-to-copy-edited-text desktop app, which runs in the background and rapidly converts spoken word into AI-transcribed, copy-edited text, pasted directly into your active application.

Download

Virus scanners on Windows can give false positives for this app, given how it uses your mic and your clipboard. If you'd like to review and compile the source code yourself, you can access it here on GitHub.

For Windows

For Mac

Download v1.0.5 for Mac (28MB ZIP) — latest
Download v1.0.2 for Mac (28MB ZIP)

You will need to download, extract, and then run the .app file.

Older versions

Text-to-Mic is open source. View the source code on GitHub.

Getting Started

Install VB-Cable. Install VB-Cable from vb-audio.com/Cable if you haven't already. This tool creates a virtual microphone on your Windows computer or Mac. Once installed, you can trigger audio to play through this virtual cable.
Add an OpenAI API key. Open the Text-to-Mic app by Scorchsoft and enter your OpenAI API key (tutorial video on setting up an API key). If you don't yet have an API key, visit platform.openai.com, sign up for a free account, set up billing and add some credit, generate an API key, and copy that key into Text-to-Mic. It isn't that expensive, but OpenAI will bill you for text-to-speech generation — check the text-to-speech and speech-to-text pricing, as well as the GPT models if you enable AI manipulation.
Set voice. Select your preferred voice for speech synthesis in the app UI.
Choose playback devices. Choose a playback device. I recommend selecting your headphones as one device and the virtual microphone (usually labelled "Cable Input (VB-Audio)") as the other.
Set the microphone to Cable Input VB-Audio in an online meeting. When you join a meeting on platforms like Teams, Zoom or Google Meet, select the Cable Input audio channel in the meeting tool's settings. This will play back any audio submitted via the tool when you hit play. Be aware that your own microphone will not work at the same time — you will need to switch back if you want to speak.
Type. Enter the text you want to convert to speech in the provided text area.
Play. Click "Play Audio" to listen to the spoken version of your text. This replays the previously generated audio clip, to prevent unnecessary use of your OpenAI API key.
Repeat what you said last. Use the "Play Last Audio" button to replay the last generated speech output.
Housekeeping. You can change the API key at any time under the "Settings" menu.
Experiment with AI manipulation. Play with the settings in "Settings > ChatGPT Manipulation" to automatically use AI to translate, change or enhance recorded or spoken words. Useful for expanding paraphrased content to increase the speed you can communicate, or to reduce vocal strain.

Example of virtual microphone selection in Google Meet:

Advanced Usage

1. ChatGPT AI manipulation

If you go to "Settings > ChatGPT Manipulation" you can turn this on and pick which model to use.

If enabled (both enabled and "auto apply to recorded transcript"), this will run your transcript through AI with the desired prompt each time you record your voice and convert it to text.

If you have enabled it but not turned on auto apply, you can trigger the action manually on any text you have put into "Text to Read", via the context menu "Input > Apply AI manipulation to text input". This only works if you have turned it on and added your API key.

2. Hotkeys

You can use a hotkey combination to trigger recording and playing of recorded text quickly. By default, the hotkeys are "ctrl+shift+0" to start the recording, then press it again to stop, transcribe and submit. "Ctrl+shift+9" stops the recording without playing it. "Ctrl+shift+8" replays the last transcribed or written text.

"Settings > Hotkey Settings" lets you customise the hotkey combinations used to trigger the above actions.

3. Presets

Click the presets button at the bottom of the app to open the presets area. You can then click a preset to add it to the "Text to Read" section, or double-click it to play it back immediately.

Once loaded for the first time, presets are stored in "config/presets.json". This means that if you close the app, you can edit them and add categories via Notepad. If you do this, please make sure you don't break or invalidate the JSON structure.

You can also edit presets from within the app, but this is limited to saving new presets to an existing category, favouriting presets, and deleting them. Any other edits must be made by editing the JSON file.

You can add a new preset by writing it into the "Text to Read" area, then selecting the category you wish to add it to at the top right of the area, and hitting save.

Practical Applications

Education. Teachers can use Text-to-Mic to give clear, consistent instruction in virtual classrooms.
Business meetings. Professionals who need to rest their voice can communicate in meetings without straining it.
Accessibility. Helps people with speech impairments communicate clearly in online meetings.
Translation. Translate your voice into another language and immediately play it back as an AI-generated voice on a virtual mic feed.
Expanding paraphrasing. Talk or type in shorthand and have AI convert it to longer form, then speak that longer version.

We created Text-to-Mic originally because a member of our team lost their voice, and we needed a simple way for them to use text-to-speech to talk with colleagues naturally — much more engaging than typing in a parallel chat channel, which often gets overlooked.

If you find Text-to-Mic useful, then please consider leaving a review. Reviews are not only lovely to read, they really help us build credibility and trust with our customers.

If you enjoy using Text-to-Mic, you might also appreciate partnering with Scorchsoft on other technology projects. We specialise in developing technically complex web and mobile applications.

Screenshots

Main UI:

Tone of voice presets manager:

AI text manipulation settings:

Frequently Asked Questions

Do I need a ChatGPT subscription to use this?

No, you do not need an OpenAI subscription to use this tool. However, you do need to set up an OpenAI key, which will charge you based on usage. The costs aren't too high for moderate use, but if you decide to use it, keep an eye on your charges for the first few days to make sure you're comfortable with the fees.

Yes. There is a simplified version that only converts text to speech using the system's built-in text-to-speech capabilities. These system voices are not as good or as advanced, but they are a useful fallback and a cost-effective option. When you load the app for the first time, if you choose not to add your API key, the application will still open and you will be able to use it. Here is a screenshot showing how this displays.

Please note that in the version without an API key you only have access to text-to-speech, not speech-to-text.

How can I find or set up my OpenAI API key?

You must sign up for an account and create a key in OpenAI's developer area. It sounds complex, but it's fairly straightforward; here is a tutorial video.

What is the difference between the GPT models in AI manipulation settings?

This setting determines which AI model is used to manipulate input or recorded text based on the provided prompt. Think of it as picking which AI brain to use.

GPT-4o mini is cheaper per word to manipulate text and is faster, but less intelligent than GPT-4o.
GPT-4o is a more powerful AI and is more likely to be able to deal with complex instructions, but it costs more per word to run and is a little slower.

We recommend trying 4o-mini first due to its speed benefits, and switching to a larger model should you find you want it to perform certain AI manipulations better.

What is the "Prompt" in the AI manipulation settings?

The prompt is the set of instructions you want the AI to use when manipulating your input or output text. The AI reads the instructions you have set in the prompt and applies them to any converted text. Here are some example prompts:

"Convert from English to Spanish"
"Expand paraphrased utterances to fully formed sentences."
"If I ask a question, reply to that question followed with a potential answer."
"Edit my input. You are a clown at an amusement park; convert to speak as this persona."
"Edit my input. You are a character in a computer game with a dark sense of humour. Convert text to speak as this persona. Remain concise"
"Copy edit my input. My mood today: upbeat, focused. Match this tone".

We recommend trying different prompts and making up your own too. You can also write much longer prompts than the examples above, if you want the AI to do something very specific. Remember to switch to a more capable model if your prompt is particularly complex or requires more accuracy. If the response replies to what you said rather than transforming it, add something like "Copy edit my input" or "Transform my input" to the prompt and that should fix it.

Remember that AI can "hallucinate" false information and give wrong answers, so evaluate responses before treating them as true.

I have ideas for new features or custom extensions that would benefit my business. Can you help me with that?

If you notice a bug or a small quality-of-life enhancement, please let us know, and we will consider implementing it in the tool for free.

We can also accommodate more substantial enhancements, such as custom extensions for business, though please be aware these are likely to carry a development charge. Please contact us to let us know what you have in mind.

Changelog

v1.4.1. The app now works without an API key, but only supports system voices and text-to-speech (no speech-to-text or other AI capabilities). System voices are also available as an option, which is useful if there are internet or connectivity issues.
v1.4.0. Lots of UI and UX improvements, added latest version checking, app now remembers previous input and output device selection.
v1.3.5. More voices added, improved presets to scale, improved keyboard shortcuts to add a cancel operation, allow banner hiding.
v1.3.0. Ability to change tone of voice. Tone of voice preset manager. Updated UI look and feel.
v1.2.0. Added presets (stored text to re-play), plus quality of life improvements.
v1.0.8. Added settings to remap hotkeys, changed .env file location to /config.
v1.0.7. Added support for hotkeys (ctrl+shift+0; ctrl+shift+9; ctrl+shift+8).
v1.0.6. Fix audio channel sample rate mismatch issues.
v1.0.5. Adds ChatGPT manipulations functionality to auto-manipulate input text.
v1.0.4. Adds input device selection option.
v1.0.3. Fixes the record button and styles better.
v1.0.2. Added Mac support, plus a record voice button (but the app crashes if audio is over around 3 seconds).
v1.0.1. First working version of the app.

Terms of Use, Disclaimer, and Licence Information

Text to Mic is provided "as is" and on an "as available" basis, without any warranties of any kind, either express or implied. Scorchsoft Ltd expressly disclaims all warranties, whether express, implied, statutory, or otherwise, including but not limited to the implied warranties of merchantability, fitness for a particular purpose, and non-infringement. We do not warrant that the software will function uninterrupted, that it is error-free, or that any errors or defects will be corrected.

Limitation of Liability

In no event will Scorchsoft Ltd be liable for any indirect, incidental, special, consequential, or punitive damages resulting from or related to your use or inability to use Text to Mic, including but not limited to damages for loss of profits, goodwill, use, data, or other intangible losses, even if Scorchsoft Ltd has been advised of the possibility of such damages.

Use at Your Own Risk

By using Text to Mic, you acknowledge and agree that you assume full responsibility for your use of the software, and that any information you send or receive during your use of the software may not be secure and may be intercepted or later acquired by unauthorised parties. Use of Text to Mic is at your sole risk.

License Agreement

This program is free software: you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

This program is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.

See the GNU Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public License along with this program. If not, see https://www.gnu.org/licenses/.

The names "Scorchsoft" and "Scorchsoft Ltd." and the associated logos are trademarks of Scorchsoft Ltd. You may use these names solely for the purpose of providing attribution, as required by the LGPL licence, and not in any way that implies an endorsement or affiliation with Scorchsoft Ltd. without explicit written permission.

DISCLAIMER: This software is provided "as-is," and any use of this software is at your own risk. For more information, see the LICENSE.md file included with this project.

Please read the full licence agreement and terms of use here before downloading or using Text to Mic (additional terms apply as described in the LICENSE.md file).

Text-to-Mic: Free AI Text-to-Speech-to-Microphone Tool (TTS & STTTS App for Windows and Mac)

In summary

Key takeaways