Last updated September 15, 2026
Link2AI receives supported public URLs only when a connected AI asks Link2AI to read them. Depending on the link, it may process public page/post text and metadata, or video metadata, captions and temporary audio needed to produce AI-readable source context.
Link2AI stores the email address used for your account, a salted password hash (never your plaintext password), account/session identifiers, plan and quota usage, and identifiers needed to manage a Lemon Squeezy subscription. To limit repeated free-trial abuse, Link2AI also uses a random first-party browser token associated with trial eligibility. Link2AI does not store card numbers.
Link2AI uses a random first-party browser token to count page views, unique visitors and a small set of product-funnel events such as starting signup or opening the Claude setup. When present, Link2AI may also store UTM campaign fields and only the referring site hostname (not the full referrer URL) so traffic quality can be measured. It does not store an IP address and does not use a third-party advertising or analytics tracker. A one-time historical backfill used existing Railway request logs to estimate visits from before first-party analytics was enabled; only anonymous synthetic visitor IDs, page paths and timestamps were imported, not raw IP addresses or user-agent strings.
Downloaded media and derived audio are temporary processing files and are deleted after the job finishes. Source-text/transcript cache is temporary and defaults to 72 hours. Completed and failed job records are retained only as minimal operational/accounting records; stored source URLs and canonical identifiers are scrubbed after the content-retention window. Unpaid pending jobs expire and have their source URL scrubbed after 24 hours.
Link2AI may send the minimum data needed to service providers used for hosting, billing, media processing, and optional email authentication. These providers process data under their own terms and privacy policies.
Link2AI does not support private accounts, paid/DRM content or bulk crawling. Scanned image-only PDFs are not OCRed in this release. Link2AI does not sell a source-content corpus and does not intentionally keep raw video, audio or scraped webpages as a content library.
Questions, support, or privacy requests: Link2AI Support ยท easonlsy1019@gmail.com