Privacy Policy
This policy truthfully describes what data QuduHub processes, where it is stored, how AI models receive inputs, and what controls you have over your account. We prioritize verified architecture over marketing jargon.
1. Operator & Platform Overview
QuduHub (accessible at qudu.in) is an intelligent workspace combining conversational AI, document exploration, and digital creation. It is operated under Pixra, an independently operated, founder-led product initiative founded by Radhe Bharadwaj.
QuduHub is currently in an active development and pre-launch preview phase (“Beta”). This document provides an accurate, evidence-based description of how the platform actually handles your information as implemented in our systems.
If you have questions regarding this policy, you can contact the operator directly at contact@quduhub.com.
2. Information We Process
We distinguish strictly between data needed to operate the core workspace and optional data you choose to provide:
A. Google Sign-In & Account Identity
When you sign in using Google OAuth 2.0 / OpenID Connect, Google shares verified profile details with QuduHub:
- Your name and email address
- Your profile picture URL
- Your unique Google Provider Account ID (
sub) - A verified authentication timestamp (
auth_time) asserting when you authenticated with Google
We never receive, request, or store your Google account password.
B. Workspace Content: Chats, Prompts & Documents
- Guest Sessions: If you use QuduHub without signing in, chat sessions and viewer files reside exclusively in your local browser cache (IndexedDB and local storage). They are not transmitted to or stored on our account servers.
- Signed-In Sessions: For authenticated users, chat titles, timestamps, and message payloads are synchronized to our cloud database and secure object storage so you can access your conversations across sessions.
- AI Memory & Preferences: When memory features are enabled, contextual facts and workspace preferences are summarized and saved to assist future AI responses.
C. Technical & Operational Data
To maintain availability, enforce rate limits (e.g., maximum 100 requests per minute to prevent denial-of-service), and debug infrastructure errors, our servers log transient request metadata including IP addresses, user agent strings, and request timestamps. These logs are used strictly for security and system stability.
D. Children's Privacy & Age Eligibility
QuduHub is a general-audience workspace designed for users aged 13 and older (or the applicable age of digital consent in your jurisdiction). We do not knowingly solicit or collect personal information from individuals under 13.
3. AI Model Providers & Third-Party Transmission
QuduHub acts as an intelligent client and workspace orchestration layer. When you submit a prompt or request AI generation, your message content is transmitted via encrypted HTTPS to external artificial intelligence infrastructure:
- Google Generative AI (Gemini): Conversation messages and title prompts are processed by Google's Gemini model family (such as Gemini Flash) to stream answers and generate session titles.
- Cloudflare Workers AI & Intent Routing: Edge routing models (such as Llama-3-8B on Cloudflare Workers AI) classify user intent to route queries between search, assistant, and support workflows.
We do not sell your personal data or conversation history to third-party data brokers. However, prompts submitted to generative models are processed by upstream AI providers under their commercial API terms. Because provider terms vary based on API tier and configuration, we do not make sweeping claims that your data is never retained or never processed for model safety checks.
Best Practice: Never submit sensitive personal identification numbers, payment card numbers, medical records, or confidential corporate credentials into chat prompts.
4. Storage, Synchronization & Sharing
QuduHub uses a hybrid storage architecture designed for speed and offline resilience:
- Client-Side IndexedDB (
QuduHubDB): Operates on your device to store compressed GZIP binary representations of chat messages, metadata, and locally cached document viewer files. - Cloud Database (Supabase PostgreSQL): Stores account profile records, session metadata, device tracking entries, and wallet transaction logs.
- Cloud Object Storage (Cloudflare R2): Stores server-synced chat sessions and user memory files, compressed with Brotli (
.json.br) under your isolated account identifier. - Public Chat Links: If you explicitly generate a public shared link for a chat session, that specific chat session's compressed content is made accessible via a unique identifier (
/api/public/chats/:sharedId) without revealing your user UUID or account credentials. Shared chats can be unshared at any time.
6. Account Deactivation vs. Permanent Deletion
We provide two distinct lifecycle mechanisms for your account:
Account Deactivation
Deactivating your account immediately marks your profile inactive (is_active = false) and records a deactivation timestamp.
All protected account APIs reject requests with HTTP 403. Your existing data is retained in a locked, dormant state.
Reversible: You can reactivate your account only by performing a fresh, interactive Google sign-in that asserts an authentication timestamp occurring after the deactivation date.
Permanent Account Deletion
Executing an account deletion request permanently wipes your profile record from the database, removes device tracking records, and deletes stored AI memory files. Synced chat sessions in cloud storage are dissociated from your identity.
Irreversible: Once deleted, your account profile and memory cannot be recovered.
Operational Note: Encrypted database backups and server log files cycle out on automated rolling retention windows.
7. Genuine Security Practices
We avoid unsubstantiated marketing claims and buzzwords. Instead, here are the concrete technical safeguards implemented in our software:
- Transport Security: All client-to-server and inter-service communications require modern TLS/HTTPS encryption.
- Token Verification: Incoming Google ID tokens are verified cryptographically against Google's public RS256 JWKS keys (validating issuer, audience, and expiry) before claims are accepted.
- Custom API Key Encryption: If you supply personal third-party AI keys, they are encrypted at rest using AES-256-GCM authenticated encryption before being saved.
- Strict Session JWTs: Internal API calls between the frontend and vault service rely on signed JSON Web Tokens using HMAC SHA-256 with strict claim verification.
- Defensive Headers & Limits: Endpoints enforce rate limiters, strict 100kb payload caps, and defensive HTTP headers (
X-Content-Type-Options: nosniff,X-Frame-Options: DENY).
8. User Rights, Inquiries & Policy Changes
You have clear choices regarding how you interact with QuduHub:
- Use as a Guest: You are free to explore and use the workspace without signing in. Your work remains on your local device.
- Data Requests: You may request access to, correction of, or permanent deletion of your stored account data by emailing us.
- Policy Updates: If our data practices change, we will update the “Effective Date” at the top of this page. Significant changes will be announced on the platform or official channels.