Free Download: Speech-to-Copy-Edited-Text Desktop App (OpenAI Whisper AI)

Quick Whisper is a free, open-source alternative to Wispr Flow for Windows and Linux. It uses AI to turn spoken audio into a clean, copy-edited transcript, then automatically pastes the finished text into whichever application you are using.

Watch a short demo of Quick Whisper running on Windows (MP4)

Speaking is generally two to three times faster than typing, so you answer emails and messages sooner. Quick Whisper runs the whole speech-to-transcription-to-copy-editing sequence in the background, so you never switch apps to have the AI tidy your text up.

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

Automatic speech-to-text conversion. Captures spoken ideas and replies as you say them, faster than typing.
Built-in AI copy editing. Refines transcriptions for clarity, readability and coherence, so you spend less time editing by hand.
Auto-paste functionality. Pastes the edited text straight into your active application, so you never have to switch apps.
Hotkey-activated recording. One shortcut starts and stops recording, so dictation fits into what you are already doing.
Customisable AI models. Pick the model that suits the job, trading cost against quality.
Adjustable settings. Auto-paste, auto-copy and AI preferences are all configurable, so the app fits how you work.

Watch the demo video linked above to see Quick Whisper in action.

Use it to compose emails, reply to colleagues, and dictate paragraphs for blog articles or social media posts. Anything that involves a lot of writing gets quicker, and the output comes out tidier.

The app is free, but running it uses your own API key. It's kind of like we've given you a free car, but you need to pay for the petrol to drive it. Although we are providing the app for free, please be aware that using an OpenAI key for transcription and AI copy editing incurs a cost. You should be mindful of these costs and keep track of them if you intend to use this tool. It is the software that we are offering for free, not the use of your API key. This isn't hugely expensive and can be mitigated by changing which model you use, and will vary by use.

Download Quick Whisper STT for Free

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.

Download the latest release

The GitHub releases page is updated more often than the list below, so it may carry a newer version than the one listed here.

For Windows

Download v2.1.2 for Windows (34MB EXE)

For Linux

Download v2.1.2 for Linux (76MB zip) (right click, save as)

For Mac

Though written in Python, this app has not yet been compiled for Mac. Please let us know if you are a Mac user who would like to use this, so we can judge demand and consider releasing for Mac too.

Older versions

GitHub source code

If Windows displays a message stating that this is an unsigned application and asks if you want to run it anyway, please be assured this is just because we haven't compiled and exported a certified copy yet. This involves more work, and for now we're releasing it for free. Perhaps we can release a certified Windows app store version later if there is enough demand.

Getting Started

Download Quick Whisper. Begin by downloading the Quick Whisper application from the download link above.
Set up your OpenAI API key. Obtain an OpenAI API key (see instructions on how to generate one) and keep it handy for setup.
Enter your API key. Open Quick Whisper and enter your OpenAI API key in the prompt that appears.
Start recording. Click the Start Recording button or use the keyboard shortcut to begin recording. Start speaking to capture your voice input.
Stop recording. To end the recording, click Stop Recording or use the same shortcut again.
Retrieve your transcription. The app will automatically process your speech and produce a refined, copy-edited transcription. You can manually copy this text from the app, or enable Auto-Copy and Auto-Paste to have it transferred directly into your active application.

Keyboard shortcuts explained:

The "Ctrl+Alt+J" keyboard shortcut will record, transcribe, and then automatically edit the transcription using AI.
The "Ctrl+Alt+Shift+J" shortcut will record and transcribe without using AI editing.

We've added both options because sometimes you might prefer a raw transcription rather than a copy-edited version.

Change Log

1.9.3 - 18/08/2025. We fixed an issue that prevented users from copying and pasting their API key (due to key binding conflicts). We also added the GPT-5 suite of models and fixed an issue where line spaces would be created at the end of pasted text. View release on GitHub
1.9.1 - 28/04/2025. Changed default key bindings to "Ctrl+Alt+J" and "Ctrl+Alt+Shift+J" to fix Windows inserting J characters erroneously. Added the latest gpt-4o-transcribe transcription model, and gpt-4.1 models for AI copyediting. Refactored code to be cleaner and more cohesive. Improved head menu usability. Implemented a system tray icon which gives the option to refresh keyboard bindings if they get lost. Added a mechanism to check and re-assign keyboard bindings every 30 seconds (toggleable). Added latest version checking with a popup prompting an update.
1.8.0 - 05/12/2024. Added "alt+left/right" keyboard shortcuts to navigate between prompts in the background (it reads their name out too). Added the ability to set custom keyboard shortcuts. Refresh key bindings when the app is maximised (sometimes it loses them when you lock your screen, so this enables a quick re-bind process).
1.7.0 - 02/12/2024. Added a Whisper input language select to prevent incorrect language detection (Whisper randomly thinks I'm speaking in Welsh sometimes). Fixed key bindings being lost; added a key binding test function. Updated the default prompt to account for different language input (translates to English by default).
1.6.1 - 25/11/2024. Code structure improvements (tidy up) and small quality-of-life improvements, addition of a cancel recording keyboard shortcut (Win+X), changes to make it ready for Mac users (pending testing), and a new prompt editing and viewing feature to enable custom editing prompts.
1.5.0 - 16/11/2024. UI improvements to enhance the visual flow and order of interface items. Buttons were also updated to be rounded, to look friendlier.
1.4.0 - 15/11/2024. Updated pop noises with different pitches so that it is clearer when a start or stop action is triggered. Added history for the current session and allowed that history to be navigated. Also added saving of the current session to a JSON file. Note that history clears when the app is closed. Changed the keyboard shortcut for transcription only to "Win+Ctrl+J". Fixed a word wrapping issue in the textarea.
1.3.0 - 14/11/2024. Replaced the AI edit tick box with two buttons and independent keyboard shortcuts, making it easier to trigger transcript only or edit only when running in the background. Added the ability to copy the last transcript and edit from the context menu. Updated the copy edit prompt to better reflect the tone of the intended use case.
1.2.0 - 13/11/2024. Main public Quick Whisper release with core functions working.

Screenshots

The main Quick Whisper app interface (v1.5.0):

Model settings adjustment:

Retry last recording:

Prompt management:

Frequently Asked Questions

What is Quick Whisper, and how does it work?

Quick Whisper is a speech-to-copy-edited-text tool that uses AI to transcribe spoken audio into polished text. The app automatically pastes the text into your active application, saving time and enhancing productivity.

What are the key benefits of using Quick Whisper?

Quick Whisper speeds up workflows by enabling faster communication, saving time on editing with built-in AI, and eliminating the need to switch between applications for transcription and copy-editing tasks.

Does Quick Whisper support multiple AI models?

Yes, Quick Whisper allows users to select from various AI models, enabling flexibility in choosing the model that best suits your performance and budget needs. At present only OpenAI models via their API are supported.

How do I set up an OpenAI key to begin?

To set up an OpenAI key, first create an account with OpenAI if you haven't already. Navigate to their playground, which serves as their developer area. Under the dashboard, locate the API keys section, where you can set up your own key and copy it for use in the application. Please be cautious not to share these API keys, as others could use your account and incur charges. This video demonstrates how to generate an API key.

How do I activate Quick Whisper's recording feature?

Quick Whisper includes a hotkey-activated recording function, which lets you start and stop recordings quickly, ensuring minimal interruptions to your workflow.

Can I customise settings in Quick Whisper?

Yes, Quick Whisper offers adjustable settings for auto-paste, auto-copy, and AI model preferences, allowing users to tailor the app to their individual preferences.

Is Quick Whisper compatible with my operating system?

Quick Whisper is designed for Windows and Linux, and should work smoothly on compatible systems, integrating into your daily tasks.

What is an API key, and what do you do with it?

The API key lets us connect to OpenAI to use their transcription and AI models within the application. You need to generate an API key, as it allows the app to use your OpenAI account. Consequently, OpenAI will bill you per use.

This setup enables us to offer the app for free, as we don't incur ongoing costs, given it utilises your API key. While the app is free, API key use is chargeable per use. We don't store the key in the cloud or have access to it on our side; it's saved only on your computer and accessible solely to you. We only use it to authenticate your access to the OpenAI API. It's crucial not to share this key with others.

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. For example, at the time of writing:

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.

The OpenAI Whisper model has been open-sourced. Why didn't you use this free version instead of using an API key that incurs charges?

Yes, you can download the Whisper model for free and run it locally, and this was an option to us when making the app; however, the model download file is quite large, often in gigabytes. The performance of running it locally depends greatly on the hardware specifications of your machine, which may result in slow operation.

Speed is crucial for the user experience of this application, so we decided to use the OpenAI Whisper API for transcription and copy editing. Although it incurs a cost, it significantly outperforms any on-device speech-to-text or copy-editing engines we've used so far. Our aim was to optimise for speed, user experience, and output quality, rather than being overly concerned about the API call cost, which is relatively minor in the grand scheme of things.

Will my speech and transcripts be used to train OpenAI's models?

No. Because we handle the transcription and copy editing using the API version of OpenAI, they have committed that data sent and processed in this way is not used to train their models.

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.

Terms of Use, Disclaimer, and Licence Information

Quick Whisper 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 Quick Whisper, 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 Quick Whisper, 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 Quick Whisper 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 Quick Whisper (additional terms apply as described in the LICENSE.md file).

Open-Source Wispr Flow Alternative: Quick Whisper for Windows and Linux

In summary

Key takeaways