Privacy Policy
This page explains how MadeMe handles your data. Core feature pages show only the brief notices needed for the action at hand.
Data we collect and use
Account and identity
Sign-in identifiers, display name, avatar, age confirmation, and confirmation that you are creating your own MadeMe, used to protect your account and establish a MadeMe that belongs only to you.
Data you provide
Identity details, experiences, views, memories, text, images, files, audio, and video, used to shape, calibrate, and manage MadeMe.
Conversations and feedback
Messages you send, MadeMe answers, supporting citations, ratings, and corrections, used to continue conversations and improve your personal version.
Security and operational records
Necessary request, upload, authorization, deletion, and security-event records, used to prevent abuse, troubleshoot, apply rate limits, and demonstrate that your choices were carried out.
MadeMe is private by default. We do not sell your personal data or use your private data for advertising.
AI and voice services
AI features are optional by default. We send only the minimum data needed for a task, and only when you enable and use that feature. You can enable or withdraw consent for all current AI features on the Control page.
Text conversation generationDeepSeek
- Purpose
- Generate MadeMe text responses grounded in confirmed personal data.
- Data sent
- Current conversation message, Confirmed memories selected for this answer, Confirmed persona and expression rules, Generated text response
- Processing region
- DeepSeek states that service data is collected, processed, and stored in mainland China.
- Retention and restrictions
- No zero-data-retention commitment is currently available. Provider-specific API terms, retention periods, and deletion capabilities must still be verified before production use.
Fallback text conversation generationOpenRouter → Azure EU → OpenAI
- Purpose
- When DeepSeek is unavailable and you have enabled the current AI service bundle, send the necessary text through OpenRouter to a fixed OpenAI GPT-5.6 Terra endpoint hosted in Azure's EU region to generate a MadeMe response.
- Data sent
- Current conversation message, Confirmed memories selected for this answer, Confirmed persona and expression rules, Generated text response
- Processing region
- Requests pass through OpenRouter before reaching the Azure EU endpoint. Model inference is fixed to that Azure EU region, but OpenRouter does not state that relay and metadata processing occur within the EU.
- Retention and restrictions
- Each request enforces zero data retention, disables data collection, and turns off prompt and response logging. OpenRouter still retains call metadata without message content. If no compliant Azure EU endpoint is available, the request fails instead of using another region.
Personal synthetic voiceElevenLabs
- Purpose
- Create and use your own synthetic MadeMe voice.
- Data sent
- Voice samples verified as your own, Voice representation and provider reference, Response text to be spoken, Generated audio
- Processing region
- Uses the global ElevenLabs API; storage or processing in a specific country or region is not guaranteed.
- Retention and restrictions
- The provider must retain the voice representation, and voice cloning does not support zero retention. Generation requests are retained according to account settings and provider policy. You can disable the voice or permanently delete the provider-held voice profile.
Image and scanned PDF recognitionOpenAI
- Purpose
- Generate drafts of OCR text, scene descriptions, and timeline clues for you to confirm.
- Data sent
- Original image or scanned PDF, Image description and capture date you provide, Recognized visible text, scene descriptions, and timeline clues
- Processing region
- Files can be stored in Singapore, but model inference is not currently guaranteed to occur in Singapore.
- Retention and restrictions
- API data is not used for training by default. Standard abuse-monitoring logs may be retained for up to 30 days. Images and files undergo CSAM scanning and may be retained for human review if flagged.
AI candidate memory extractionDeepSeek
- Purpose
- Generate candidate memories from one reviewed source for you to confirm individually.
- Data sent
- Text-derived content from one reviewed source, Model-generated candidate memories and source evidence excerpts
- Processing region
- DeepSeek states that service data is collected, processed, and stored in mainland China.
- Retention and restrictions
- No zero-data-retention commitment is currently available. Provider-specific API terms, retention periods, and deletion capabilities must still be verified before production use.
Audio and video transcriptionOpenAI
- Purpose
- Generate an editable transcript draft visible only to you and requiring your confirmation.
- Data sent
- Original audio or MP4 container, Transcript text, Fixed file type and model parameters
- Processing region
- Uses a Singapore API endpoint. Data can be stored in Singapore, but model inference is not currently guaranteed to occur there.
- Retention and restrictions
- Current transcription endpoint documentation states that no application state is retained and that zero data retention can be requested. Singapore-region use still requires approval for modified abuse monitoring or zero data retention.
Conversation voice inputOpenAI
- Purpose
- Turn one voice input into editable conversation text before it is sent.
- Data sent
- One temporary conversation recording up to 60 seconds, Transcript text, Fixed file type and model parameters
- Processing region
- Uses a Singapore API endpoint. Data can be stored in Singapore, but model inference is not currently guaranteed to occur there.
- Retention and restrictions
- Current transcription endpoint documentation states that no application state is retained and that zero data retention can be requested. Singapore-region use still requires approval for modified abuse monitoring or zero data retention.
If a provider or processing method changes materially, we will update this page and ask you to confirm again when required.
Retention, deletion, and export
- Source data remains in private storage until you delete individual items or permanently delete your account.
- Pausing MadeMe does not delete data; it only stops answers and invitation access.
- When you delete source files, parsed text, memories, voice representations, or your account, the interface clearly explains the scope of the deletion.
- You can export a copy of your data on the Control page and permanently delete your account and associated data in Account settings.
Your choices and rights
- View, correct, export, or delete data you provide.
- Enable or withdraw AI and voice services without affecting your ability to manage saved source data.
- Disable or permanently delete an AI voice and delete the corresponding source voice samples.
- Revoke invitations, pause MadeMe, or permanently delete your entire account.
MadeMe is only for users aged 18 or older creating one for themselves. If this policy changes materially, we will update the date and request confirmation again when required.