
EmailOps
Chat with your inbox - locally, privately, free
Details
- Follow on
- Categories
- AIProductivity & Notes
- Use Cases
- Email ManagementEmail Security
- Target Audience
- Mac UsersML EngineersSaaS Founders
- Pricing
- Free
- Platforms
- Desktop
Discovery signals
How AI and people discover EmailOps on PeerPush
About EmailOps
EmailOps is a free, open-source email client for desktop with AI built in - but unlike most AI email tools, the model runs entirely on your machine. Your email is never sent to a cloud service to be summarized, sorted, or drafted. Connect Gmail, Outlook/Microsoft 365, or any IMAP account (ProtonMail Bridge, Fastmail, iCloud) and: See everything in one unified inbox — all your accounts, one view Chat with your inbox in plain language — "what did the landlord say about the deposit?" Auto-classify mail and drive smart filters Get AI-drafted replies without your messages leaving your laptop How it stays private: the AI is an embedded model running in-process via llama.cpp — by default there are zero calls to any cloud AI. Mail is stored in a local SQLite database, and OAuth tokens live in the macOS keychain. The only way anything reaches a hosted model is if you deliberately switch the provider yourself. The entire codebase, including the local/network boundary, is open source — read it or build from source if you don't want to take our word for it. Status: early alpha (v0.6.2). Runs on Apple Silicon Macs (M1 or newer). Windows and Linux are on the roadmap. Expect rough edges — and please report them. Use cases Privacy-first email management with AI assistance Searching and querying your inbox in natural language Automatic email classification and smart filtering AI-drafted replies processed entirely on-device Managing multiple accounts (Gmail, Outlook, IMAP) in one unified inbox Offline-capable AI email triage
Screenshots
Reviews (2)
Average 5.0 out of 5
Based on 2 reviews
Top ! 👏🏻


Comments (1)
I built EmailOps because every AI email tool I tried wanted to ship my mail to its servers to do anything useful. So I built a client where the AI conveniences run locally, and open-sourced it.