CME has processed a real iCloud inbox with roughly 600,000 messages spanning about 15 years. Not a demo account. Not a benchmark with seeded data. A real inbox that belonged to a real person who had never seriously cleaned it. That experience shaped how the product handles large inboxes — and it's worth explaining what that actually required.
If your inbox has tens of thousands of messages and cleanup feels impossible, this post explains the architecture, what the scan covers, and what you'll actually see before anything is moved or deleted.
Why Large Inboxes Are Different
At a few hundred emails, manual cleanup is tedious but doable. At ten thousand, it's overwhelming. At 600,000, it's not a project you can reasonably approach without tooling. The categories you'd want to clean — newsletters, automated notifications, old bulk mail, expired alerts — are distributed across years of history in ways no search filter can isolate cleanly.
The problem isn't just volume. It's that most of the messages have no obvious label. "Newsletter from a company you no longer remember" looks like a normal email. Finding all of them requires pattern recognition across senders, domains, frequency, and content — not a keyword search.
How the Architecture Handles Scale
When CME runs a sweep on a large inbox, several things happen in sequence:
Parallel IMAP workers fetch message envelopes from the provider simultaneously. Envelopes contain the headers — from, to, subject, date, message ID — without the full body. Pulling envelopes in parallel rather than sequentially is what makes the initial scan fast even for very large inboxes.
Adaptive page sizes reduce round trips. Rather than fetching a fixed number of messages per request, the scan adjusts based on response latency, which matters when network conditions vary across a long-running job.
A single O(n) classification pass runs over the fetched envelopes. Each message is assigned to a category — newsletters, notifications, receipts, conversations, automated mail — based on sender patterns, domain signals, and header metadata. This pass is CPU-bound and measured in seconds even at large scale.
Cleanup simulation converts those categories into an action plan: how many messages would be archived, what percentage of inbox volume each category represents, and which specific messages fall into each bucket.
The full provider scan is network-bound and can take several minutes for very large inboxes. Classification is fast. That distinction matters: if you're watching a progress indicator, most of the wait is IMAP latency, not analysis.
What the 600k Sweep Actually Found
In the iCloud inbox described above, the largest cleanup opportunities were newsletters and automated mail. Notifications — shipping updates, service alerts, bank statements, app digests — made up a significant share of the total volume. Old bulk messages from vendors and services the user had long since stopped using were another major category.
What wasn't obvious until the sweep ran: how little of a 600,000-message inbox is actually irreplaceable. The messages a person would genuinely regret losing — real conversations, important receipts, documents, project threads — represent a small fraction of total volume. The simulation surfaced that clearly, before anything was touched.
Plan Caps vs. Architecture Ceiling
The 600k story is architecture proof, not a promise about every plan. CME's pricing tiers have different cleanup action limits:
- Free — up to 25,000 emails per sweep, simulation-only; no executable cleanup actions
- Pro ($9/month) — 25,000 cleanup actions per month, write actions enabled
- Power ($25/month) — 150,000 cleanup actions per month, priority queue
- One-Time Sweep ($19, one-time) — up to 250,000 emails, single inbox
For a 600k inbox, the free tier will classify the full inbox and show you the simulation, but executing cleanup across all of it requires a paid plan or the One-Time Sweep. The scan itself runs regardless of plan. What's gated is the number of cleanup actions you can execute per month.
Simulation Before Anything Moves
No matter how large the inbox, the sequence is always: sweep → classify → simulate → review → approve → execute. You see what would happen, including sample messages from each category, before CME touches anything. Cleanup doesn't start until you say so.
That design is especially important for large inboxes, where the consequences of a mistake are harder to unwind. Archive is the default recommendation over delete. Reversible actions first, destructive actions only after explicit review.
Example Prompts for Large-Inbox Sweeps
- "Sweep my inbox and show the classification breakdown."
- "Which cleanup strategies reduce active inbox load the most?"
- "Simulate archiving newsletters and notifications — show me what would be affected."
- "Show me what will be preserved before any cleanup runs."
FAQ
Can every plan scan a 600,000-email inbox?
The sweep and classification can run across plans. What differs is how many messages each plan scans and whether cleanup actions can execute. The free tier caps each scan at 25,000 emails and executes no cleanup actions. For a full 600k cleanup, you'd need to split the work across eligible paid passes; compare the current One-Time Sweep and Power limits before starting.
How long does a large-inbox scan take?
The provider scan is network-bound — fetching envelopes from a 600k inbox via IMAP takes several minutes depending on the provider. Classification runs in seconds once the envelopes are fetched.
What's the biggest risk with large-inbox cleanup?
Acting before you understand what you're cleaning. The simulation step exists precisely to prevent that. You should not approve a cleanup action until you've reviewed sample messages from the category and you're confident the rule is right.
Does CME store my email content?
CME does not maintain a secondary archive of your inbox. It processes messages to classify and surface them; metadata and feature-specific retention are handled as described in the privacy policy.
Is archive always safer than delete for large inboxes?
Yes. Archiving removes messages from your active inbox while keeping them searchable and recoverable. Permanent deletion should be reserved for categories you've inspected and are confident about.
If your inbox has grown to a size where manual cleanup stopped being realistic, a sweep report is the right first step. See the full tools list, compare plans, or start with the free email cleaner to run a classification without committing to any cleanup actions.