Privacy

Last updated November 1, 2026

EverPosted is built so your archives never leave your computer. This page lists everything that does leave it, who receives it, and what they keep.

What stays in your browser

• The exports you pick. The page reads them in your browser and never uploads them.
• The identifiers you list for replacement: names, handles, emails, phone numbers and the rest.
• Your report, which the page builds in your browser and you save as a file.
• Your report code, kept in this browser's local storage. The page deletes it on your first visit after it expires, and clearing the site's data deletes it sooner.
• The two name-detection models, about 290 MB, kept in your browser's cache after you first build a report.

What leaves your computer

When a free scan ends, the page sends one request to everposted.com, and that day's count of scans goes up by one. The request carries no cookie, no id and nothing from your files. It can carry short codes from the list below, for what the scan could not read or why it stopped. A code names a platform at most, never a file or anything in one.

When you build a paid report, the page first sends your report code alone to our scoring service at score.everposted.com, which answers how many items the code has left. Then it sends short redacted excerpts to the same service, with your report code. Before anything is sent, the page replaces your names, handles, emails, phone numbers and the other identifiers you listed, and other people's names it finds. Each excerpt is at most 600 characters, and a run sends at most 1,500 of them. With each excerpt goes a fixed set of facts: the platform, the kind of item, its visibility, its year, whether you wrote it, who can delete it, whether the account is publicly tied to your real name, and which sensitive topics it touched. The service checks every excerpt again for leftover emails, phone numbers, links and long numbers, and refuses a request that holds one. When the run ends, the page sends one more request to everposted.com, like the scan's, with codes for how the run ended and what it could not read.

The codes a count can carry

everposted.com refuses a count with any code not on this list.

• tgz: a picked file was a .tgz or .tar archive, which the page cannot open.
• not_zip: a picked file was not a zip.
• unrecognized: a picked export matched none of the six platforms.
• google_html: a Google export held its activity in HTML, which the page does not read.
• google_no_activity: a Google export had no My Activity folder.
• facebook_html: a Facebook export was in HTML, which the page does not read.
• nothing_read_ and a platform (facebook, x, reddit, google, discord or chatgpt): nothing in that platform's export could be read.
• unread_ and a platform: that platform's export held files the page neither reads nor knows to skip.
• stopped_crash: the scan or run stopped because the tab ran out of memory or crashed.
• stopped_error: the scan or run stopped on an error.
• complete, partial or failed: how a paid run ended.

When you pay, you leave everposted.com for Stripe's checkout page. When Stripe sends you back, the page passes Stripe's id for that checkout to our scoring service. The service asks Stripe whether the payment went through, whether it was refunded or disputed, and whether the billing address is in the US. Stripe's answer includes your name, email and address. The service reads only what those checks need and keeps none of it.

Who receives it, and what they keep

• Our scoring service forwards the excerpts to Anthropic and returns the scores. It stores no excerpts and no scores. For each report it keeps how many items and requests its code has used, and deletes that a day after the report's last code expires. For each payment it confirms, it keeps the code's expiry and limit under a hash of Stripe's checkout id until the code expires or 30 days pass, whichever is later. We can match that hash to the payment in our Stripe account. Its logs hold the code's report id, item counts, the model's token counts, status codes and timing. Cloudflare keeps those logs for up to 7 days.
• Anthropic runs the language model that scores the excerpts, in the United States. Anthropic does not train its models on them. It deletes them within 30 days, except content its safety systems flag, which it may keep for up to two years.
• Stripe processes your payment and keeps your name, email, billing address and payment details under Stripe's privacy policy. We can see those details in our Stripe account to handle refunds and questions, and we copy them nowhere else.
• Cloudflare hosts everposted.com and the scoring service. Like any host, it sees your IP address and basic request details. It also gives us daily totals of visits, which hold no personal details, and stores our counts of scans and runs with their codes. The scoring service uses your IP address briefly to limit how often a payment or a report code can be checked.
• Email. If you write to hello@everposted.com, we keep the conversation as long as we need it to help you.

What we do not do

No cookies, no analytics or tracking scripts, and no ads. We do not sell or share personal information. No person at EverPosted reads your archives, the excerpts or your report.

Your choices

• Clear this site's data in your browser to remove your report code and the cached models.
• To ask what we hold about you, or to have it deleted, email hello@everposted.com. For most people that means the Stripe record of a payment and any emails.

Changes and contact

We update this page whenever the code changes what leaves your computer, and the date at the top shows the latest version. EverPosted is a service of JBowz Holdings LLC, a Texas limited liability company. Questions go to hello@everposted.com.