Privacy Policy

Last updated: May 7, 2026

tl;dr

We only collect what we need to run groups, voice notes, voice feedback, and light usage signals.
When you send a voice memo, we transcribe it with AI (OpenAI). That audio is sent to OpenAI for transcription so others can read what you said.
To play pronunciation audio we send plain text through OpenAI first, ElevenLabs only if OpenAI speech is unavailable. No replay of your own clip there.
Your account rides on Supabase in the United States. We never sell your data.
Ping noah@language-soup.com for access, corrections, exports, deletes, or questions.

What We Collect

Profile information: Name, profile photo (optional), and languages you are learning.
Voice messages: Audio you send in group chats.
Voice transcription: When you send a voice memo, we transcribe it with AI. We send that audio to OpenAI (Whisper) to create a text transcript others can read. The recording itself stays in your group on our servers.
Voice feedback (optional): If you ask for speaking feedback, we also send audio to Groq for analysis. We process it in real time for that feature.
Pronunciation playback: When you tap to hear feedback read aloud we send only the phrase text through OpenAI text-to-speech. If OpenAI speech is unavailable in our setup we route that same text to ElevenLabs. They return audio only.
Usage data: App activity such as features used and typical session length, so we can improve the experience.

Camera Permission

We only ask for camera access if you want to take a profile photo. You can skip this and upload from your photo library instead.

How We Use Your Data

To run language learning groups and daily challenges.
To connect you with other learners.
To improve the app experience.

Third Parties

When you send a voice memo, we send that audio to OpenAI (Whisper) for transcription. If you use optional speaking feedback, Groq may also process audio for that feature. They work on our behalf and do not own your data.

For pronunciation playback we send text to OpenAI text-to-speech first. ElevenLabs can fill in as a backup if OpenAI speech is not configured. They also work for us and only see the text we pass in for that sound bite.

User data and authentication are managed through Supabase, which stores data on servers in the United States.

We do not sell your data to any third party.

Each vendor runs its own privacy policy: OpenAI, Groq, ElevenLabs, and Supabase.

Lawful Basis

We process your data to provide the Language Soup service. This is called contract performance.

We also process some data to improve the app experience. This is called legitimate interest.

Data Storage

Your data is stored securely on Supabase servers. We never sell your personal information.

Data Retention

Profile and usage data are kept while your account is active and deleted when you close it.

Transcription runs when you send a memo or tap to transcribe. Optional feedback audio is kept only as long as needed to return your result.

Your Rights

You can reach noah@language-soup.com anytime to ask about access to your personal data, fixes, exporting what we hold, or closing your account and deleting associated data.

If you live in the EU, you can rely on formal legal rights under privacy law for the same kinds of requests, using that same inbox.

Contact

Questions? Email us at noah@language-soup.com.