About Sayelle AI
Sayelle's tutor and roleplay characters are artificial intelligence systems, not people. The tutor generates replies and learning feedback using a language model downloaded to your device.
What happens on your device
Speech recognition, tutor response generation, and text-to-speech run locally. Raw microphone audio is not uploaded. Saved text transcripts and learning progress sync to Sayelle's backend when connected. Reliable recovery and cross-device access require linking the account with Sign in with Apple on iPhone or Sign in with Google on Android. See the Privacy Policy.
Important limitations
• Safety instructions and filters reduce risk but cannot guarantee that every response will be suitable or correct.
• Explicit first-person self-harm or immediate-harm language is checked on-device before generation. When detected, Sayelle bypasses the normal AI reply and shows a fixed supportive response directing the learner to local emergency services and a trusted person.
• Roleplay characters are fictional and do not know your real circumstances.
• Medical, pharmacy, legal, financial, insurance, and emergency scenarios are language practice only—not professional advice.
• Use made-up personal details in practice. Do not enter real phone numbers, addresses, account numbers, health details, legal case details, passwords, or secrets.
• For a real emergency or important professional decision, contact an appropriate qualified local professional or emergency service.
Feedback and reporting
You can report an unsafe or seriously incorrect response through Sayelle's reporting controls or from Profile. Reports are voluntary and should not include unnecessary personal information. You can also contact us through the Support page.