SmrutiCortex Privacy Policy

Last Updated: March 4, 2026

1. Data Collection

SmrutiCortex does NOT collect, transmit, or store any personal data externally. All data processing happens locally on your device.

2. What Data is Stored Locally

The extension stores the following data in your browser's local IndexedDB:

• Browser history URLs (obtained via Chrome History API)
• Page titles
• Visit timestamps and visit counts
• Optional metadata (keywords, descriptions) extracted from pages you visit
• Optional AI embeddings (if semantic search is enabled) — generated locally via Ollama, never sent to any cloud service
• Extension settings and preferences

3. Where Data is Stored

All data is stored exclusively in:

• Browser IndexedDB (local database on your device)
• Chrome Storage Local API (settings only)

No cloud storage or external servers are used.

4. Data Transmission

SmrutiCortex does NOT:

• Send any data to external servers
• Use cloud storage or sync services
• Transmit browsing history anywhere
• Include analytics or tracking tools
• Connect to any cloud AI service (OpenAI, Google, Anthropic, or any other)
• Send your search queries, browsing history, or page content to the internet for AI processing

Optional Features:

• Favicon Loading: When enabled, fetches site icons via Google Favicon API (can be disabled in Settings). This is the only network request SmrutiCortex makes.
• Local AI (Ollama): All AI features connect strictly and exclusively to Ollama running on your local machine (typically at http://localhost:11434). Ollama is a separate open-source application that you install and control — it runs AI models entirely on your device. SmrutiCortex never connects to any cloud or remote AI service. Whether Ollama itself accesses the internet depends on your own Ollama configuration, but SmrutiCortex only communicates with your local Ollama instance.

AI Features — Privacy Details:

• AI Keyword Expansion: Your search query is sent only to your local Ollama instance to generate synonyms. The query never leaves your device over the internet.
• Semantic Search (Embeddings): Page titles and URLs are sent only to your local Ollama instance to generate vector embeddings. These embeddings are stored locally in IndexedDB. No page content or embeddings are transmitted externally.
• No Cloud Fallback: If Ollama is not running or not installed, AI features are simply disabled. SmrutiCortex never falls back to a cloud AI service — it falls back to keyword-only search, which requires zero network access.
• No Model Downloads: SmrutiCortex does not download AI models. You must manually pull models using Ollama's own commands (ollama pull model-name:tag). SmrutiCortex only uses models already available on your machine.

5. Data Retention

Data is stored indefinitely until you:

• Manually clear extension data via Settings ("Clear & Rebuild")
• Uninstall the extension (automatic cleanup)
• Clear browser data

6. User Control

You have full control to:

• Delete all extension data anytime via Settings
• Disable metadata extraction
• Disable favicon loading
• Disable AI features
• Add sensitive sites to blacklist (no indexing for banks, password managers)
• Uninstall the extension (removes all data)

7. Permissions Used

The extension requires these permissions for core functionality:

• history: Read browser history for indexing and search
• bookmarks: Read bookmarks for unified search
• storage: Save settings and indexed data locally
• tabs: Query active tab and send messages to content scripts for inline overlay
• alarms: Keep service worker alive and schedule background indexing updates
• <all_urls>: (Optional) Extract metadata from pages user visits (local processing only)

8. Third-Party Data Sharing

SmrutiCortex does NOT sell or transfer user data to third parties.

No browsing history, URLs, or personal information is shared with any external service, company, or individual.

9. Security

All data is stored locally using browser's secure storage APIs (IndexedDB, chrome.storage). Only you have access to your data. The extension cannot access data from other extensions or websites you haven't visited.

10. Open Source

SmrutiCortex is open source. You can review the code at any time:

https://github.com/dhruvinrsoni/smruti-cortex

11. Changes to Policy

Updates to this privacy policy will be posted at:

• https://dhruvinrsoni.github.io/smruti-cortex/privacy.html
• GitHub Repository Documentation

12. Contact

Questions or concerns about privacy?

• Open an issue: GitHub Issues
• View documentation: GitHub Repository

13. Compliance

This extension complies with:

• Chrome Web Store Developer Program Policies
• User Data Privacy requirements
• Limited Use disclosure requirements