We actually read our messages.

TurboMail has three plans. Student ($0.99/mo) includes every individual feature - full local AI, unlimited accounts, AI chat & insights, and calendar sync. Professional ($2.99/mo) adds priority model updates, advanced AI actions, web search, text-to-speech, and early access to new features. Enterprise is custom-priced and adds team features like centralized billing, shared AI model config, admin controls, and dedicated support. All plans get a free trial period at launch.

Your emails stay on your device — always. TurboMail stores a local SQLite cache of your emails for fast search and offline access. The AI model processes emails locally without any network calls. We have no server infrastructure that touches your email data.

Yes! TurboMail auto-detects your hardware and recommends the right model. On older or CPU-only machines, we suggest Phi-3 Mini (~2.4 GB, runs on 8 GB RAM). It's slower than GPU-accelerated models but fully functional. On Apple Silicon, even an M1 MacBook Air delivers impressive speeds via Metal/MPS.

Gmail personal and Google Workspace accounts work via OAuth. Microsoft 365 and Outlook.com work via OAuth too. For corporate Microsoft 365 accounts, your IT admin may need to approve the app - standard procedure for any desktop email client.

The app itself is ~185–190 MB. AI models are downloaded separately on first setup: the smallest (Phi-3 Mini) is ~2.4 GB, the recommended Qwen 2.5 1.5B is ~2 GB, and the powerful Llama 3.1 8B is ~4.7 GB. You only need one model at a time, and you can change or remove them anytime.

Yes. TurboMail works as a regular email client without any AI model. Email sync, composing, sending, HTML rendering, calendar sync, and all core features work independently. You can add AI anytime from the Settings → AI Model screen.

For bugs: open an issue on our GitHub. For security vulnerabilities, email support@turbomail.ai with "Security Disclosure" in the subject line - we respond within 48 hours and will credit you in the changelog.