Privacy Policy

Last updated: July 2026

Important Demo Notice: Archivist is a Retrieval-Augmented Generation (RAG) demonstration platform. We prioritize your document security. Uploaded materials are processed solely to facilitate your immediate chat session interaction. You can skip uploading a document and just chat immediately with the bot based on our default document. View current system default file →

1. Information We Collect

Archivist only handles data strictly required to demonstrate our document chat capabilities. This includes:

Uploaded Files: Text files, PDFs, or CSVs you intentionally upload to query.
Chat Prompts: Text queries you submit directly to the conversational AI bot.
Session Cache: Temporary browser memory tokens to maintain your login and active chat stream.

2. How Your Data Is Processed

Your documents undergo a localized process called vector embedding creation. This text breakdown feeds our Retrieval-Augmented Generation pipeline.

We do not utilize your custom files, conversational queries, or contextual data prompts to train public machine learning systems or baseline LLM models.

3. Data Retention and Deletion

Because Archivist operates primarily as an architectural tool demonstration, data persistence is intentionally short. Document vectors and chat interaction histories are purged automatically from our servers when you signout.

4. Third-Party API Processors

To handle conversational responses, text segments of your document are selectively parsed to secure Large Language Model processing endpoints. These external processors are tightly bound by commercial privacy agreements which restrict them from retaining your session information.

Contact and Questions

Have inquiries about this sandbox privacy framework? Contact the demo administrators directly at Chibuz.com.