Last updated: 20 December 2025
This page lists the third-party AI service providers ("AI subprocessors") that process personal data on behalf of Stratara Ltd in connection with the Sotto couples messaging app.
Sotto's default partner-to-partner messaging is designed so that your raw message history and full-resolution media live on your devices, and messages/media are relayed end-to-end encrypted between partners.
This subprocessor list focuses on the specific cases where Sotto uses third-party AI APIs to provide features like insights, transcription, and discreet notification filtering.
We only use AI subprocessors where necessary to deliver Sotto's core AI features. We do not use advertising networks or third-party marketing analytics SDKs in Sotto.
For more information about how we collect and use data (including encryption, retention, and user controls), please see the Sotto Privacy Policy.
Role: Primary AI provider for Sotto's nightly relationship insights and AI features, including:
Models (typical): GPT-5.1, GPT-5.2, GPT-5-nano, Whisper
Data processed (by feature):
What we do not intentionally send: your account password, encryption keys, or your full message history archive. We aim to avoid sending your direct contact details (like email) to AI providers as part of AI prompts. (Note: providers may still receive technical metadata required to operate their services, such as IP address and request logs, under their own policies.)
Locations: United States and other regions as described in OpenAI's documentation/policies.
Notes:
OpenAI processes data as a service provider when we call their API. We minimise the content we send (bounded chunks, not full archives) and minimise exposure time. We may use available account-level or API-tier privacy controls where offered by OpenAI, but OpenAI's own terms and policies also apply during processing.
More information:
Role: Lightweight AI provider for specific real-time or low-latency tasks, such as:
Models (typical): Llama 3.1 8B
Data processed (by feature):
What we do not intentionally send: your encryption keys, full message archive, or full-resolution media files. We aim to avoid sending direct contact details (like email) in prompts. (As with most API services, Groq may receive technical metadata like IP address and request logs under its policies.)
Locations: United States and other regions as described in Groq's policies.
Notes:
These calls are typically short-lived and scoped to the specific feature (e.g., preview classification). We aim to minimise what is sent and avoid storing raw inputs on our systems longer than needed.
More information:
This page lists AI subprocessors that may receive message content or derived content (like transcripts or descriptions) for AI processing.
Other third parties may be involved in operating Sotto (for example: Apple services for push notifications, App Store purchases, and platform functions). Those are described in the Sotto Privacy Policy and are not necessarily listed here unless they act as AI processors of message content.
For Sotto, we do not use:
We may add, remove, or replace AI subprocessors as Sotto evolves. When we do so, we will:
If a change materially affects your rights or how your personal data is processed, we will also reflect it in the Sotto Privacy Policy and, where required by law, notify you in-app or through other appropriate means.
If you have questions about any AI subprocessor listed here, contact us:
Privacy: [email protected]
Support: [email protected]
Related documents:
Privacy Policy |
Terms of Service