Privacy Policy — DocsChat

1. Information We Collect

DocsChat provides conversational AI document intelligence. When you use our platform, we collect account information (name, email address, password hashes), billing identifiers, and the documents you upload for training (PDFs, DOCX files, text snippets, and crawled website URLs).

2. Document Indexing & Vector Search

Uploaded documents are chunked and converted into vector embeddings. These vectors are stored strictly in dedicated MongoDB Atlas database collections mapped to your tenant workspace. We do not use your proprietary documents or customer conversations to train third-party foundation models.

3. Multi-Model AI Orchestration

Inference requests are processed via your selected provider (Google Gemini, OpenAI, or Anthropic). Only the minimal relevant document context retrieved via cosine similarity is transmitted in transient inference prompts.

4. Data Retention & Deletion

You maintain complete control of your data. When you delete a document source or agent from your dashboard, all underlying document chunks and vector embeddings are permanently expunged from MongoDB Atlas.

5. Contact Our Privacy Team

If you have any questions or data protection requests (GDPR, CCPA), contact us at privacy@docschat.ai.