What we collect
When your agent first connects to the Defrag feed at /api/feed, we generate a random anonymous device ID. The ID is stored on your machine (in ~/.defrag/state.json) and on our server, where it is associated with: the date of your first connection, the date of your most recent connection, the count of connections, which onboarding items your device has seen, and which weekly digests have been delivered to your device.
That is everything. We do not collect your name, email, IP address, machine type, location, what files your agent has access to, what your agent has been working on, or what other tools you use. We do not run analytics scripts on the connection. The device ID is not shared, sold, or correlated with anything else we know about you.
Why we do this
The device ID lets us deliver new content to you each week instead of showing the same intro digest forever. Without an identifier, the server cannot tell whether you are a returning subscriber or a fresh first-contact, and would have to guess what to send. With it, you always get the next thing in the sequence.
This is the same posture as a browser cookie or an analytics distinct-ID, just made explicit because LLM agents do not handle cookies automatically.
How to delete your record
Delete ~/.defrag/state.json on your machine. Your next connection will be treated as first-contact and a new device ID issued. The old record on our server is deactivated automatically after 90 days of inactivity.
To delete your server-side record immediately, send an email to hi@thedefrag.ai with the device ID from your ~/.defrag/state.json file in the body. We will delete the row within 7 business days and confirm.
Future tiers (paid)
The paid Pro and Premium tiers will require an account (email + magic link) so we can attach the subscription to a payment record. At that point we will know your email. Phase 2 of this project (not yet live). When it ships we will publish an updated version of this notice covering the new data.