Privacy Policy

SignetKit (“we”, “us”, or “our”) respects your privacy and is committed to protecting the personal data and documents you entrust to us. This Privacy Policy explains what information we collect, how we process it, and the legal and technical safeguards we apply. Please read it carefully.

1. Data we collect

• Account information: name, email address, and profile avatar provided through Google OAuth sign-in.
• Documents and files: PDFs, images, resumes, contracts, letterheads, and any other files you upload, generate, or sign through the Service.
• Usage data: IP address, browser type, timestamps, feature usage, and error logs, used only for security, debugging, and service improvement.
• Payment data: subscription tier and billing identifiers processed by our payment provider. We do not store full payment card details.

2. How we process your documents

Our document pipeline is designed so that raw sensitive content is never retained in plaintext after ingestion:

• Upload: Files are sent directly from your browser to Cloudflare R2 object storage via a short-lived, signed HTTPS URL. Our application servers never see the uploaded bytes.
• Rest encryption: R2 encrypts objects at rest using server-side AES-256 managed by Cloudflare.
• Semantic indexing: a background worker extracts text from PDFs, splits it into chunks, computes mathematical vector embeddings, and then encrypts every text chunk using AES-256-GCM with a unique 12-byte nonce before writing it to our database.
• Selective column encryption:the database stores only the ciphertext, nonce, and authentication tag in dedicated binary columns. There is no plaintext “content” column. The encryption key is loaded from a secure environment variable and never leaves the application layer.
• Pre-retrieval isolation:vector similarity queries are scoped to your authenticated user_id at the database level. Other users’ chunks are excluded before any distance calculation occurs.

3. AI and large language model disclosures

SignetKit provides an optional document assistant that uses retrieved chunks to answer questions. We disclose the following regarding our AI architecture:

• Ephemeral processing:decrypted text exists only in volatile memory for the duration of a single inference request. It is not written to disk, cached in a database, or retained in application logs.
• No model training: your documents, queries, and generated answers are never used to train, fine-tune, or improve any foundation or third-party AI model.
• No retention of prompts or context: after an answer is returned, the decrypted context is immediately discarded. We do not keep a history of your questions.
• Third-party LLM provider:inference may be served by a self-hosted or third-party LLM provider configured by the deployment administrator. Data is transmitted to that provider only to generate the immediate answer and is governed by the provider’s data processing terms.

4. Password-protected documents

We do not currently support client-side decryption of password-protected PDFs. If you upload a password-protected file, you must unlock it before upload using a PDF tool you trust. The file stored by SignetKit is the unencrypted version you upload. You are responsible for ensuring that you have the legal right to unlock and upload any protected document.

5. Third-party sub-processors

We use the following sub-processors to provide the Service. Each is contractually bound to process data only as instructed:

• Cloudflare, Inc. — object storage (R2) and presigned upload URLs.
• Google LLC — OAuth authentication and Google account profile data.
• Razorpay Software Private Limited — payment processing and subscription management.
• Ollama / configured LLM provider — ephemeral document question-answering inference.
• PostgreSQL hosting provider — encrypted metadata and vector index storage.

6. Your rights

Depending on your jurisdiction, you may have rights to access, correct, delete, restrict, or export your personal data. You can delete individual documents, your account, or request a full data export by contacting us at [email protected].

7. Data retention and deletion

We retain your documents and account data for as long as your account is active. When you delete a document, its R2 object, encrypted chunks, and associated metadata are removed promptly. Backup snapshots may retain encrypted fragments for a limited retention window. Aggregated, non-identifiable usage statistics may be retained for analytics.

8. Changes to this policy

We may update this Privacy Policy from time to time. We will post the revised version on this page and update the “Last updated” date. Continued use of the Service after changes constitutes acceptance.

9. Contact us

For privacy-related questions, data subject requests, or security disclosures, please contact [email protected].