tt Multi-Model AI Client FAQ

tt provides integrations for model providers including OpenAI, Anthropic, Google, xAI, DeepSeek, Kimi, Doubao, Qwen, Zhipu GLM, and MiniMax, plus managed platforms such as AWS Bedrock, Azure AI, Google Vertex AI, and Alibaba Cloud Model Studio. Account eligibility, regions, quotas, and available models still depend on each service. See the AI providers and hosting platforms page; the current configuration list in the app is the source of truth.

Model requests travel directly from your device to the AI providers you configure. When Online Search is active, a derived query may be sent to teramoby's Cloudflare-hosted search service; the current Android version first asks the summarizer whether current web information is needed. See the Privacy Policy for platform behavior and the complete data flow.

Open tt, go to Settings → Model Configuration, select your provider, paste your API key, and save. Keys are stored securely in the device's system secure storage and are never written to ordinary files as plain text.

In a standard chat, tt sends your question to every currently configured service with a complete model and credential, preserves each completed response, and passes them to your chosen summarizer. With only one configured service, that single response is shown directly. The result is still model-generated content, so important claims should be verified.

Team Mode gives every currently configured service with a complete model and credential a shared discussion. They can respond to the user and to one another, add new information, challenge assumptions, or stop responding when they have nothing new to add. An @ mention asks one model to respond first without excluding the others. The discussion ends automatically when every participant is quiet. Learn how Team Mode works.

Use standard chat for a quick answer, side-by-side comparison, or a synthesized result. Use Team Mode when a question benefits from critique, trade-off analysis, or several models building on one another over a shared discussion.

These official pages cover several commonly used services. See each provider's documentation for account eligibility and current credential requirements:

• ChatGPT (OpenAI)Get API Key →
• Claude (Anthropic)Get API Key →
• Grok (xAI)Open xAI Console →
• Gemini (Google)Get API Key →
• DeepSeekGet API Key →
• AWS BedrockGet API Key →
• Doubao (VolcEngine Ark)Get API Key →

Yes. Conversation history is stored as app data on your device. Your operating system may include that local app data in a system backup or device transfer according to your platform settings. On Android, conversation history and settings can be included, while encrypted API keys are excluded.

API keys are stored in the device's system secure storage and are not written to ordinary files as plain text. A key is used to authenticate directly with the provider you configured; it is not sent to teramoby or unrelated services.

tt does not send usage analytics, advertising identifiers, or full conversation history to teramoby. Requests you initiate still go to the AI providers you configure. When Online Search is active, a derived query may go to teramoby's Cloudflare-hosted search service. Online Search is enabled by default in current releases and can be turned off in Settings. See the Privacy Policy for platform behavior, endpoint, caching, and third-party processing details.

tt is free to download and use. You only pay for the API calls you make to the AI providers you choose — billed directly by those providers at their own rates. tt takes no cut.

tt is available for Android, iPhone, iPad, and Mac. Download Android directly from the website; on a supported App Store storefront, search for teramoby and select “teramoby tera (tt).” The app is not currently listed in the mainland China App Store.