Cred Count

What changed in Version 2

Published 4 October 2026.

Version 1 of this policy was published on 4 August 2026. Version 2 describes Cred Count 1.5.0, the version going into public testing, and the changes to our server and our suppliers since Version 1. We've also rewritten the policy to be shorter and plainer.

The Privacy Policy is the full notice, and it's the one to read. This page is a guide to what's different in Version 2, and why, so you don't have to compare the two versions yourself.

Save a copyPDF59 KBMarkdown8 KB

1. Smart Import's AI moved to the European Union

Before: the sample of your spreadsheet that Smart Import sends went to Google's Gemini on a global endpoint, which meant Google could process it in any country. Version 1 said so in §15: "there is exactly one route to a US machine, and it is Smart Import."

Now: Smart Import runs exclusively on Google's European Union multi-region, where Google publishes a commitment that processing stays inside the EU. Everything about you that we handle is now stored and processed in Europe, so that sentence in §15 is gone.

Why the EU, and not one country. The model Smart Import uses is only offered on Google's US and EU multi-regions, so the EU is the narrowest choice available. The UK recognises the EU as giving data adequate protection, so this raises no transfer question.

What stayed the same. Smart Import sends the same sample as before: the first 100 rows, up to 4,000 characters. Only where it goes changed.

2. The AI ride matching is gone

Before: Version 1's §7 described a second AI feature. When the app couldn't match a ride name you'd typed ("BTR", or a misspelling), you could send the ride and park names to the AI to work out what you meant, with your device ID.

Now: that feature has been removed from the app. §7 describes Smart Import alone, and §4 lists two things that carry your device ID instead of three. Matching the names you type happens on your device.

3. Google Drive: three permissions became two

Before: Cred Count asked your Google account for three permissions: its private backup folder, reading a spreadsheet, and listing every spreadsheet in your Drive, so the app could show you a menu to choose from. Google classes that last one as restricted, its most serious tier, because it reveals the name of every spreadsheet you own.

Now: two permissions, and Google classes both as non-sensitive, its lowest tier. Google's own file picker shows you your files on Google's side and hands back only the one you choose, so the app never needs to list your Drive.

Two more changes came with it:

§3, §8 and §13 also now point to drive.google.com/drive/settings → Manage apps → Cred Count, where you can disconnect the app and delete its hidden data from your own Google account, even after you've uninstalled. §8 now says that linking backup signs you in with Google, which gives the app your basic Google profile and email address, and lists everything a backup holds.

4. Google keeps less of your spreadsheet

Before: Google logged Smart Import requests for up to 55 days, as a matter of course, to detect misuse. Google's AI platform also caches inputs in memory by default, for up to 24 hours.

Now: nothing is logged unless Google's automated safety checks flag a request. A flagged request can be kept for up to 90 days so a person can review it, stays in the EU region, and isn't used for training. We've also switched off the in-memory cache on our project. Google doesn't count caching as keeping data, but a default nobody chose isn't a good reason for your spreadsheet to sit anywhere.

5. Server backups and submissions have time limits

Now: our server's backups are deleted after 30 days, and §14 has a row for them. That's inside the 90 days a submission's device ID and comment are kept.

Submissions have an outer limit too. A submission's device ID and comment are erased 90 days after we accept or reject it, and now also 12 months after you sent it, whatever has happened to it (§9).

6. Deleted things are erased after six months

Before: Version 1 said the app marks things you delete as deleted, so you can undo a mistake, and didn't say for how long.

Now: after six months, the app erases them completely (§3).

7. We're on the ICO's register

Clever Little Goose LLC is registered with the UK's Information Commissioner's Office as a data controller, reference ZC213564. §1 carries it, §13 gives it as the reference to quote if you complain, and §15 uses it to show that being a New Mexico company doesn't put us outside UK data protection law.

8. A paragraph we removed from §15

Version 1's §15 said that Google and our email provider are global groups, and that their data processing agreements let group companies outside Europe reach data when supporting the service needs it.

It was in the wrong section. §15 is about where your data goes: your spreadsheet and your IP address. Someone helping Zoho fix a mailbox isn't touching either. Anything our signed agreement with Zoho says about who can reach email sent to us belongs in §10, the section about email.

9. Crash reports from Apple and Google

Before: §12 said "No crash reporting. If it crashes, we find out because you tell us."

That's still true of the app: nothing in Cred Count reports a crash. We now publish releases to the Google Play Store and Apple App Store. Apple and Google collect crash reports themselves, if you've allowed it, and show developers what crashed, on which model and operating system, and where in the code.

Now: §12 says there's no crash reporter in the app, and a new §12.1 explains what Apple and Google collect, under whose settings, and what a report does and doesn't contain.

10. Other changes in Cred Count 1.5.0

11. More detail in the policy

What didn't change

Reading the difference yourself

Version 1 is still published, exactly as it was, in the archive: the page, the Markdown and the PDF. We don't edit a published version. When the policy changes, we publish a new one and explain it on a page like this.

Questions about any of this: privacy@credcount.app.