About VoiceMerge
VoiceMerge is made by a small independent team based in the United States, and operated by Nameverse, Inc.
Why it works on your device first
Dictation hears everything: the email to your doctor, the half-formed idea, a client's name. We think the right default place for that audio is the device you spoke into. Macs and iPhones are fast enough to turn speech into text without a server, so VoiceMerge does that there. A cloud model is something you switch on for a job that needs it, not the price of using the app.
Every dictation says which path it took. If something left the device, you chose it, and you can see it.
How it's paid for
The apps are free, with no trial and no feature gates. On iPhone, optional credits pay for Claude by the request. We charge for the part that costs us money to run, and nothing else.
What we don't do
- No analytics, tracking or ads, in the apps or on this site.
- No storing the text of requests that pass through our service.
- No selling or sharing your data for marketing.
What it's built on
VoiceMerge stands on open-source software and openly licensed speech models: Parakeet from NVIDIA, Whisper from OpenAI, the pyannote Community-1 speaker models, FluidAudio and WhisperKit. The licenses page credits each of them, with their licenses; the apps show the same list under Acknowledgements.
Contact
Help: [email protected]. Privacy: [email protected]. Security reports: see the security page. The legal details are in the privacy policy and the terms.