Conversation mode turns a single device into a live two-way interpreter: two people talk, each in their own language, and both sides appear translated on the one screen.
How far it reaches depends on how many audio sources you can give it, and that differs by platform.
| Languages | Quality & Medical | |
|---|---|---|
| iPhone | 10 | Not in Conversation |
| Windows, Mac — one source | 10 | Not in Conversation |
| Windows, Mac — microphone and computer sound | All 50+ | Yes |
One source — 10 languages, on every platform
Pick a single source and any two of 10 supported languages: English, Spanish, French, German, Italian, Portuguese, Russian, Japanese, Dutch, and Hindi.
The sources you can choose from depend on the device:
- iPhone: the microphone, or your device's audio
- Windows and Mac: the microphone, your system audio, or one chosen app
StreamVox listens to that one stream, detects which of the two languages every spoken line is in, and translates it into the other.
Both directions appear in one feed. Each block is labeled with its direction (English → Español, Español → English), and each speaker gets their own color.
This is the setup for:
- A job interview or client meeting across a table (microphone)
- A reception desk or front office serving foreign visitors (microphone)
- A family dinner where not everyone speaks the same language (microphone)
- A stream or podcast where the host and guest speak different languages (device audio)
- A press conference where questions and answers don't match (device audio)
- A game voice chat where your team talks in two languages at once (device audio)
Both sources at once — every language, plus Quality and Medical (Windows and Mac)
Turn on the microphone and computer sound together, and StreamVox no longer has to guess which language belongs to which side. The microphone is one channel, computer sound is another, so there is nothing left to detect between them.
That removes the 10-language ceiling: every language StreamVox supports (see supported languages) becomes available in Conversation mode.
Quality and Medical modes work here too, though they are normally solo-source only. Each side even keeps its own recognition settings, tuned independently instead of shared.
Reach for this when the person across the table speaks a language outside the curated 10, or when the conversation calls for Quality or Medical mode's precision.
This two-source setup is a Windows and Mac feature. iOS gives an app one microphone stream at a time, so on iPhone Conversation always runs on a single source: 10 languages, without Quality or Medical. Everything else here works the same.
What it is not
Conversation mode is not the way to make the other side of a Zoom call read your language — your subtitles are visible only on your own screen.
For calls, each participant runs StreamVox on their own device with their own language pair. That setup scales to any number of participants.
See Does StreamVox work with Zoom, Teams, and Discord? for that flow.
Switching between Translate and Conversation
Each mode applies its own settings on a fresh start: press Stop, switch the mode, and start again. It takes a few seconds.
Switching modes — or changing a language — clears the current text from the screen. Copy anything you want to keep first, using the Copy button on either pane.