Immediate source deletion
The private source PDF is deleted immediately after OCR completes.
Security and privacy
Bank statements are sensitive. Tennanova limits how long source files and extracted results exist, restricts result access to the verified account owner, and keeps financial data out of analytics and advertising events.
These controls describe the current production design. No internet service can promise absolute security, so Tennanova also keeps the processing window narrow and asks you to review and download results promptly.
The private source PDF is deleted immediately after OCR completes.
Successful extracted results automatically become inaccessible after two hours.
Sensitive job payloads use AES-256-GCM encryption while stored in Cloudflare D1.
Result and edit endpoints verify the signed-in Clerk account that created the job.
Production traffic is protected in transit with HTTPS and TLS.
Statement and transaction data are excluded from product analytics and advertising events.
Data lifecycle
The PDF is streamed into private Cloudflare R2 storage after account and security checks.
The document is sent to the OCR provider inline or through a short-lived signed URL for larger supported files.
Tennanova deletes the R2 object immediately after OCR. Scheduled cleanup and storage lifecycle rules act as backstops.
The normalized extraction is encrypted in D1 and can be accessed only by the account owner through authorized endpoints.
The result becomes inaccessible two hours after successful processing and is then physically removed.
Tennanova uses specialist providers for authentication, hosting, OCR, billing, product analytics, and advertising measurement. Each receives information relevant to its role. The privacy policy contains the complete processing and location disclosures.
| Purpose | Provider | Boundary |
|---|---|---|
| Authentication | Clerk | Account identity and verified sign-in |
| Hosting and temporary storage | Cloudflare | Site delivery, private PDF storage, encrypted temporary results |
| Document reading | Zhipu AI and Z.ai | Statement content required to perform the conversion |
| Billing | Stripe | Checkout and subscription management, not full card details to Tennanova |
| Measurement | PostHog and Reddit | Allow-listed product events and advertising events, excluding statement data |
Read the full privacy policy for processing locations, retention details, analytics, advertising, and contact options.
Answers
The source PDF is stored privately only while OCR is performed and is deleted immediately after that step. A 15-minute cleanup and a one-day storage lifecycle rule provide additional backstops.
A successful result becomes inaccessible two hours after processing and is then physically removed. Download the CSV or Excel file before the temporary result expires.
No. Result and edit requests verify the signed-in Clerk account owner. A random job identifier does not grant access by itself.
No. Statement contents, filenames, balances, account identifiers, transaction descriptions, and transaction values are excluded from product analytics and the Reddit Pixel. The pixel measures page visits and selected conversion events.