Narracore Screenplay Formatter
Screenplay text you submit is processed in memory to produce structure blocks and the PDF. It is not written to persistent storage, not logged, and never used for model training or analytics datasets.
The original video is never uploaded. It stays on your machine, where your AI watches it. Only the structured breakdown text your agent writes is sent to us; it exists only while the request is being processed and is not persisted as database records. It is not logged and never used for training. PDF reports generated by your request are stored temporarily and deleted after 24 hours.
PDFs are stored behind unguessable, signed links (32-hex-character random IDs with HMAC signatures) and are automatically deleted after 24 hours. Backup routines do not include them.
We keep minimal operational logs per request: timestamp, language track, block/page counts, shot/scene counts, duration, and outcome. No script content, no breakdown content, no license keys (only the last 4 characters for support), no personal data beyond standard server diagnostics.
The web server keeps an access log (request time, path, status, user agent, IP address) which rotates out automatically at a fixed size cap and is never archived elsewhere. Query strings are not logged — so a license key passed to the checkout page or a PDF download signature never reaches the log. The public page-view counter on the home page is aggregate only: it stores daily totals and a per-day salted hash of the visitor address, never the address itself.
There are no accounts. Paid credits attach to a license key (e.g. NRC-XXXX-…) that you receive at purchase. The key's plain text is shown once; we store only its hash.
Payments are processed by PayPal. We never see or store your card data. See the pricing page for refunds (full refund within 7 days if unused; failed formats are never charged).