Is Fezola safe to give access to my mail?
In short. Fezola is built so your assistant reads and drafts but does not send on its own: replies are created as drafts, sending is off until you enable it per mailbox, and you can revoke access at any time. You still decide what to connect, so check each permission.
Updated on
Details
Replies are saved as drafts in the mailbox that received the message. Direct sending is opt-in, mailbox by mailbox. Mail content, such as an email, is treated as data and never as instructions. Attachments over 10 MB, and images over 3.5 MB, are not read. Hosted connectors cannot reach addresses on your own computer or private network. For Google, Fezola asks for full mailbox access, which it needs to use IMAP with OAuth. You can disconnect any source in Fezola and revoke access at your provider.
Keep reading
More from the docs
- How do I connect Fezola to Claude?Add Fezola to Claude as a custom connector: Customize, Connectors, Add custom connector, paste your Fezola URL and sign in. Plans and steps, checked Nov 2026.
- How do I connect Fezola to ChatGPT?Add Fezola to ChatGPT as a custom MCP server: Settings, Plugins, plus, Add custom MCP server, paste your URL. Business, Enterprise and Edu. Checked Nov 2026.
- How do accounts and sources work in Fezola?In Fezola each mailbox or tool account is a named source, free up to 2 sources. Results name their account, and one URL serves Claude and ChatGPT.
- What is Fezola Memory?Fezola Memory is one shared memory for Claude and ChatGPT: short statements about you, filed by project, on by default, and you can view, edit and erase it.