That is the whole design, not a setting. This page explains exactly what is kept, what is never kept, and the few places where something outside Seera's control still sees your connection.
Effective 28 July 2026 · Seera is in development; this describes how it is built and how it will operate at launch.
Everything in the first list is a number. None of it can answer who, or what they searched for.
People use Seera to check strangers — a buyer, a seller, an account that messaged them out of nowhere. A record of who checked whom would be worth more to the wrong person than it would ever be worth to us. So it is never written down.
The practical version: there is no database of visitors, no log of requests, and no way to reconstruct a search after it happens — including for the person who runs Seera.
You type a username. Seera asks Reddit's official API for that account's public profile, comments, and posts, calculates the summary in memory, sends it back to your browser, and discards it. Nothing is written to disk at any point.
Because nothing is cached, every analysis reflects Reddit as it is right now. If a Redditor deletes something, it disappears from Seera's view immediately — there is no copy.
Running a free service means knowing whether ten people or ten thousand are using it, and whether it is broken. That comes from counters that go up by one, never from records of what happened.
Being honest about privacy means naming the parts that are not ours.
Seera is hosted on Cloudflare, which serves the site and protects it from attacks. Like any host, Cloudflare processes connection data including your IP address in order to do that. This is outside Seera's control and is covered by Cloudflare's privacy policy. Cloudflare's Turnstile check also runs on the deep analysis button to keep automated abuse from running up the bill; it is designed to work without tracking you across sites.
Account data comes from Reddit's official API, under their developer terms. Seera reads public data only — no private messages, no moderation data, and it never posts, votes, or messages anyone. Seera does not identify you to Reddit; the request comes from the server, not from your browser.
If you run the optional deep analysis, the public Reddit content that was just fetched is sent to Anthropic's API to write the summary, then discarded. Your identity is not part of that request. The content is not used to train any model, and Seera keeps no copy of the input or the output.
Seera reads what Reddit already publishes and keeps none of it, so there is nothing held about you to correct or delete. Two limits are permanent parts of the design rather than promises:
If this policy changes, the date at the top changes with it. Anything that would weaken the guarantees above would be a change to how Seera is built, not a quiet edit to this page.
Questions about this policy: privacy@seera.fyi