A calmer inbox starts in your terminal

A little order.
A lot more
headspace.

Meet Inboxed. Organize your email with AI running on your computer, review what it suggests, and make room for what matters.

Open source · Gmail & IMAP · Local AI

An ink drawing of letters in a paper tray, with a leafy sprig growing beside them

A little room to breathe.

Built for your terminal.Powered by local Ollama models.Made to keep you in control.

A WORKSPACE WITH SOME BREATHING ROOM

Your email.
Your model.
Your call.

Categories, topic tags, and a review queue, brought together in one terminal workspace. Start with a small scan and take it from there.

inboxed / personalLOCAL

$ inboxed scan --limit 3

Scan complete. Your mailbox is unchanged.

MESSAGECLASSPROPOSAL
Field NotesThis week’s readingNewsletterArchive
The Paper ShopYour receiptReceiptLabel
Studio NorthFriday project notesWorkLabel
3 proposals ready for review0 applied

Illustrative preview. Example messages, real workflow.

LESS SORTING. MORE LIVING.

From inbox to in order.

A simple rhythm for the email you want to keep, the clutter you want to clear, and everything in between.

01 / CONNECT

Your mail, connected.

Sign in to Gmail in your browser, or connect an IMAP account.

Your provider. Your account.
02 / CLASSIFY

A place for everything.

Run a local Ollama model to classify messages into your categories and optionally suggest topic tags. Scans save proposals for review.

AI running on your computer.
03 / REVIEW

You make the call.

Review, edit, approve, or reject suggestions. Apply labels, archive, or move messages to Trash when you’re ready. Automation is an explicit choice.

Review first. Apply when ready.

A PERSONAL WORKSPACE

Your email deserves
a little privacy.

Read our privacy policy

Local processing

Email content goes to your local Ollama instance for classification and optional tags. Inboxed has no hosted mailbox backend.

Transparent storage

Tokens, scan history, message metadata, and previews stay in your local app data. You control their retention.

Open to inspection

Read the code, understand the permissions, and choose your own categories.

A FRESH START, ONE COMMAND AWAY

Meet your new
email routine.

You’ll need Python 3.13+, uv, and a running local Ollama instance with a compatible model. The README covers model setup and configuration.

Read the setup guide
GET INBOXED
git clone https://github.com/marvalkrystof/inboxed.git
cd inboxed
uv sync --extra tui
uv run inboxed init
uv run inboxed tui

Open the Accounts tab to connect, then scan and review.

BUILT IN THE OPEN

Questions, ideas, or a stubborn inbox?

Inboxed is maintained by Kryštof Marval. Find documentation, report a problem, or help shape the project on GitHub.

Get in touch