Your Inbox.
Private AI.
On-Device.
Inboxed is the first private email client powered by a Local Private LLM. Achieve zero inbox with on-device AI summarization—all without your data ever leaving your device.
Features.
Designed for the privacy-conscious professional.
Local Private LLM
The first private AI email client for Mac. Runs 7B+ parameter models (Llama 3, Mistral) entirely on your Apple Silicon GPU via llama.cpp. No cloud API, no latency, no data leaks.
No Tracking
Zero telemetry. Zero analytics. A truly private email client. Your data is your business. We rely on your device's power, not our servers.
Zero Inbox
AI-powered email client for the modern era. Automatically summarize emails and categorize receipts. Clear your inbox in seconds — even as an offline email client on a plane.
Direct Fetch
On-device intelligence with direct IMAP/Google fetch. Credentials stored in macOS Keychain. Data stored in local SQLite. The ultimate secure email client.
Engineered for Silicon.
Built on a secure, memory-safe backend. The app binary is just ~10MB — compared to 200MB+ Electron apps. Uses ~80MB RAM at idle vs 500MB+ for typical email clients.
We use llama.cpp optimized for Metal to run 7B+ parameter models directly on your GPU. Inference at ~30 tokens/sec on M1, faster on M2/M3/M4.
Your credentials stay in the macOS Keychain. Your data stays in a local SQLite database. 0 bytes sent to external servers. No analytics. No telemetry.
Fair Pricing.
No subscriptions. No hidden costs.
Standard
- — All AI Models
- — Unlimited Accounts
- — Community Support
Pro Lifetime
- — Everything in Standard
- — Priority Support
- — Early Access Features
- — Support Development
See how we compare
"The email client for the AI era."
Stop sending your personal data to the cloud to get smart features. Inboxed brings the power of Large Language Models directly to your device.