Many researchers and professionals cannot upload sensitive data to third-party servers. Mercuryscribe keeps you in control of who can see your data. If you have more data than your personal machine can handle, you can rent a virtual server from a cloud provider like AWS or Digital Oceanand install this software there, ensuring that your data is not shared with third-party transcription services. Services like these do not access data on their virtual machines.
Yes. Mercuryscribe runs on your machine or your private server. It pulls models from HuggingFace, but no audio, transcripts, or metadata are sent to external services. The software is open-source, so you (or programmers that you trust) can verify that no data is sent externally.
If having the data on your computer withing HIPAA or GDPR compliance standards, Mercuryscribe is too. You should consult with your compliance team about your specific requirements and implementation.
Yes! Mercuryscribe is completely free and open-source under the MIT license. You can use it for personal, academic, or commercial purposes without any fees or licensing costs.
Mercuryscribe is free. If you need help making it work or deploying it widely, enterprise support covers additional services like priority support, custom installations, SLAs, and pre-configured solutions. Contact us for enterprise support options.
Do you need output in a specific format or integration with your existing tools? We can discuss custom development options. Would you like to automate transcription for large batches of files as they come in from your research team? We can help with that too.
Yes! Literate Computing is owned and operated by a former academic and instructional technology professor. We offer installation help, team training, workshops for research groups, and private server setup with GPUs for large-scale processing. Learn more about our services.
Absolutely! We can handle the entire setup process, including Docker configuration, performance optimization, and integration with your existing workflows. This is especially popular with research institutions and enterprises. If you would like help with installation via screen sharing, we can handle that too.
Yes! We can deploy Mercuryscribe on your private cloud infrastructure with GPU acceleration, automated workflows, and enterprise-grade security. We are happy to work with your IT team to ensure a smooth deployment, providing them with as much are as little help as you need.
On the Intel Linux boxes and Intel and M2 Macs that I have tested, it takes about as long as the audio is, so a ten-minute audio/video file takes about ten minutes to transcribe.
Windows, macOS, and Linux are all supported. The Docker-based installation makes setup consistent across all platforms. The Python package should work on all platforms, but requires `ffmpeg`, so Brew is required on Mac and WSL is required for Windows. Getting all of the Python dependencies installed and working can be tricky, so we recommend the Docker installation for most users.
Running the Docker container (after you have installed Docker) requires pasting a single command (I wouldn't call it "code"). If you would like it installed on a private server with password protection (virtual servers from places like Digital Ocean are an option), we offer professional installation services for organizations that prefer hands-off setup.
Accuracy depends on audio quality and language, but Faster-whisper Base model typically achieves 80%+ accuracy on clear English audio. Performance is comparable to leading cloud services while maintaining complete privacy. A future version will allow using larger models for improved accuracy (but reduced speed).
Definitely! We welcome pull requests to add features and fix bugs. For custom featurescontact us for an estimate.
Yes! Support for MAXQDA is the default. We plan to support other tools as demand for them grows. We can also develop custom integrations with your favorite tools, content management systems, or analysis software.
This project was started to support a single research team. Considerable time has been invested to make it useful to a broader audience. If you find it valuable, please consider supporting ongoing development and maintenance through one of the options below.