PureBox.ai

Sender history is a strong signal but it breaks in one spot that's bitten me: a reply to a thread I started, from an address I've never opened before, looks statistically identical to a cold email. I run a lot of outbound and the messages I most need to see are exactly those first replies from strangers. Does PureBox weight "this is in a thread I sent" above sender history, or is the classification mostly per-message?

Shaked, the origin story is the strongest part of this page, and review-first reads like a decision you made for yourself rather than for the launch.One case nobody raised: a small clinic inbox is full of patients writing about symptoms and lab results, so your first scan is processing regulated data. The question there is not retention (Chalermpon and Gal already covered that one), it is who counts as the processor. That inbox owner needs a signed agreement covering you and whichever model provider you call, otherwise they cannot connect at all.Is the provider behind the sorting one you could sign that with? Solo therapists have the worst inboxes I have seen and almost no tool they are allowed to touch.

the "nothing moves until you approve" framing plus the 30 day undo window is the right way to earn trust with something touching a real inbox. one thing I didn't see covered - if someone connects, tries it out, then revokes Gmail access later, is there anything left behind on your end from the scan (cached content, summaries, sender history), or does disconnecting clear it out completely?

This is a very timely and useful solution.I used to create new accounts so I wouldn't lose anything important, and I had to unsubscribe from unnecessary newsletters.

@shaked_i Congrats on the launch! Since this is scoped to Gmail right now, is multiple account support on the roadmap, like handling a personal and a work inbox in one place?

the 30 day trash grace period is a smart touch, most people won't trust an AI email tool until there's a real undo. one thing I'm curious about - once something gets filed to Archive, does it stay just as searchable as normal gmail archive, or does the LLM sorting change how findable stuff like old invoices/contracts is later?

Congrats on the launch! The three-bucket plus reasoning approach is a smart way to keep trust intact, most inbox cleanup tools break it the moment they act silently on something that mattered.Before I'd trust it with my own Gmail, I'm curious what happens to an email's content after it goes through the model to generate the category and explanation, whether it's discarded right after the suggestion is shown or logged somewhere on your end. That distinction would decide whether I connect a real inbox to it.

I honestly have around 5k of unread emails, but I like the fact that nothing gets deleted automatically. How long did it take before you were comfortable relying on your own product?

This is clever. Does the AI learn from rejections too, or just from what you approve?

Hey I really want to use this! My email box is sooooooooo full! I am afraid I will delete something important.

Review-first is the right call. Most inbox AI I've tried loses my trust the first time it quietly archives something that actually mattered. How do you handle the gray-zone stuff, like a cold pitch that's genuinely relevant versus noise? Is the review queue learning per-user, or are the rules more global at the start?

I would never want an AI cleaning my inbox without showing me what it is doing first. The review-first approach feels like the right way to build this. Congrats!

For about two years, opening Gmail made my chest tighten.

It wasn't really the volume - it was the fear that somewhere under thousands of unread newsletters, promos, and receipts there was a bill I'd missed, or a message from an actual person I'd left on read for three weeks. So I did what a lot of us do: I stopped looking. Which, obviously, made it worse.

I tried the usual stuff. "Inbox zero" lasted a weekend. Filters I set up once and never touched again. A couple of those "auto-clean" apps that delete things in the background - which somehow made the anxiety worse, because now I didn't even know what was gone.

What I actually wanted was almost embarrassingly simple: something to sort the pile, tell me what mattered, and let me be the one to hit delete.

So I built it. It's called PureBox.ai. You connect Gmail through Google's OAuth (you grant access, it never sees your password), and it uses an LLM to read each email and sort everything into three piles: Attention (probably needs you), Archive (safe to file), and Trash (safe to bin). Every suggestion comes with a plain-English reason - like "newsletter, unopened for 8 months" - so it's not a black box. Nothing moves until you approve it, and because it all happens in your real Gmail, anything trashed just sits in the bin for 30 days. I built the review-first part first, honestly, because I didn't fully trust my own tool.

The thing that surprised me: the relief didn't come from the inbox being empty. It came from not being afraid to open it.

It's free to run on a sample of your inbox, so you can see whether it's actually useful before paying for anything. And happy to answer anything about how it handles your data - that's a fair question for anything you'd hook up to your inbox.