xtrace-sdk  by XTraceAI

Encrypted vector database for private AI memory and search

Created 2 months ago
840 stars

Top 41.9% on SourcePulse

GitHubView on GitHub
Project Summary

XTraceAI/xtrace-sdk provides an open-source SDK for secure vector databases and AI agent memory, addressing data privacy concerns. It ensures documents and embedding vectors remain encrypted on the user's machine throughout storage and search operations on the XTrace server, enabling AI capabilities without compromising sensitive information.

How It Works

The SDK employs client-side encryption: documents are embedded locally, then vectors and text are encrypted using Paillier homomorphic encryption and AES-256, respectively. The secret key stays local. The XTrace server processes only ciphertexts, performing searches and returning encrypted results without ever accessing plaintext, ensuring zero-knowledge operations.

Quick Start & Requirements

Install via uv pip install xtrace-ai-sdk or uv pip install "xtrace-ai-sdk[embedding]". Requires Python 3.11+. Obtain an API key and org ID from app.xtrace.ai for the rate-limited free tier. Full documentation is at docs.xtrace.ai. The CLI (xtrace-ai-sdk[cli]) simplifies setup and search.

Highlighted Details

  • End-to-End Encryption: Data and vectors remain encrypted locally, never exposed in plaintext to the server.
  • Zero-Knowledge Server: Operates exclusively on ciphertexts, ensuring data confidentiality during search.
  • Dual Modules: Includes x-vec for encrypted vector search and x-mem for encrypted AI agent memory (upcoming).
  • Verifiable Security: Offline tests (pytest tests/x_vec/) validate cryptographic primitives used in the SDK.

Maintenance & Community

The README provides no specific details on contributors, sponsorships, or community channels (e.g., Discord/Slack). Assessment of community health and maintenance activity requires further repository investigation.

Licensing & Compatibility

Licensed under Apache 2.0, permitting commercial use and integration into closed-source applications.

Limitations & Caveats

The x-mem module for encrypted agent memory is currently in development ("coming soon"). The free tier is rate-limited, potentially affecting high-volume usage.

Health Check
Last Commit

3 weeks ago

Responsiveness

Inactive

Pull Requests (30d)
0
Issues (30d)
0
Star History
696 stars in the last 30 days

Explore Similar Projects

Starred by Elie Bursztein Elie Bursztein(Cybersecurity Lead at Google DeepMind), Andrew Trask Andrew Trask(Research Scientist at Google DeepMind), and
1 more.

tf-encrypted by tf-encrypted

0%
1k
Framework for encrypted machine learning in TensorFlow
Created 8 years ago
Updated 1 year ago
Feedback? Help us improve.