We're thrilled to announce the official launch of the Qeltrix Organization on GitHub, alongside the release of comprehensive integration test results for our V1 Proof-of-Concept. We're sharing these results with the community to demonstrate our progress and gather feedback.
Official GitHub Organization: https://github.com/Qeltrix
Qeltrix Repository: https://github.com/Qeltrix/Qeltrix
Test Results Repository: https://github.com/Qeltrix/test-poc-1
Important Context: This is a Proof-of-Concept—not even pre-development stage yet. We're simply validating code to prove the concept works. V1 is basic and serves purely as a technical validation exercise.
What is Qeltrix?
Qeltrix is an approach to data obfuscation and streaming that combines modern cryptography, compression, and parallel processing into a unified system. At its core, Qeltrix implements Content-Derived Obfuscation—the cryptographic key is mathematically derived from the data itself, creating near-perfect statistical randomness while simplifying key management.
Current Status: Qeltrix V1 is a Proof-of-Concept. While still very basic, the PoC demonstrates something special: its architecture successfully validates the fundamental principles of high-throughput parallel processing combined with cryptographic obfuscation using industry-standard building blocks (LZ4, ChaCha20-Poly1305, HKDF).
Note: V1 implements parallel processing only during the packing phase. Unpacking currently runs sequentially, which is reflected in the throughput measurements.
Why This PoC Matters
Despite being an early-stage proof-of-concept, V1 showcases a robust architectural foundation:
- Parallel Processing: ProcessPoolExecutor enables high-throughput compression across multiple cores
- Modern Cryptography: ChaCha20-Poly1305 (AEAD) ensures both confidentiality and integrity
- Efficient Compression: LZ4 algorithm provides rapid data reduction
- Secure Key Derivation: HKDF generates deterministic, cryptographically secure keys
- Streaming Architecture: Minimal memory footprint handles multi-GB files without loading them into RAM
This isn't just prototype code—it's a working demonstration that the architectural vision is sound and achievable with proven technologies.
V1 Integration Test Results
Our test suite validates two critical metrics: obfuscation quality (entropy) and streaming performance (throughput).
| Test ID | File Type | Mode | Ratio (%) | Input Entropy | Output Entropy | Throughput (MB/s) | Status |
|---|---|---|---|---|---|---|---|
| Highly Compressible Text (8MB) | Text | two_pass |
0.43 | 3.24 | 7.99 | 44.8 | ✅ PASS |
| Low Compressibility Binary (8MB) | Binary | two_pass |
100.0 | 8.00 | 8.00 | 1.8 | ✅ PASS |
| Large Binary File Stress | Binary | single_pass_firstN |
100.0 | 8.00 | 8.00 | 17.5 | ✅ PASS |
| Mixed File Type | Mixed | two_pass |
66.8 | 6.55 | 7.99 | 36.0 | ✅ PASS |
| Low Compressibility Binary (4MB) | Binary | two_pass |
100.0 | 8.00 | 8.00 | 9.5 | ✅ PASS |
Key Findings
1. Cryptographic Obfuscation Works Perfectly
Output entropy consistently reaches ~8.0 bits/byte—the theoretical maximum for random data. The standout example: highly compressible text with just 3.24 bits/byte input entropy transforms to 7.99 bits/byte output. ChaCha20-Poly1305 completely destroys statistical patterns, making the encrypted data indistinguishable from random noise.
2. Parallel Architecture Delivers Speed (Packing Only)
The parallel processing design excels with compressible data during packing, achieving 44.8 MB/s throughput on highly compressible text. V1 currently implements parallelization only in the packing phase—unpacking remains sequential, which explains the throughput numbers shown in the test results.
3. Mode Selection Impacts Performance
Two key derivation modes offer different trade-offs:
-
two_passmode: Derives keys from entire file content, slower on incompressible data (1.8 MB/s) -
single_pass_firstNmode: Derives keys from file headers only, significantly faster on high-entropy files (17.5 MB/s)
This validates the architectural flexibility for different use cases.
Understanding Qeltrix's Vision
Essential Reading: Qeltrix: One Vision, Multiple Proofs-of-Concept
A common misconception: V1 through V5 aren't separate tools—they're incremental proofs-of-concept demonstrating different aspects of one unified architecture. Think of them as chapters proving that each piece of the overall vision works, all built on trusted cryptographic standards.
Get Involved
Qeltrix is a by-the-community, for-the-community project. The author explicitly states: there is no guarantee of updates from the author himself. The project's future depends entirely on community interest and contributions.
Licensing:
-
test_qeltrix.pyin the test-poc-1 repository is licensed under GPLv3 - We welcome developers, security researchers, and enthusiasts to:
- Explore the V1 PoC code
- Review the test results
- Suggest improvements
- Report issues
- Submit pull requests
Even small contributions can significantly impact the project's evolution.
Join us in building the future of content-derived data protection.
Happy coding!
The Qeltrix Team
Links:
- Organization: https://github.com/Qeltrix
- Qeltrix Repository: https://github.com/Qeltrix/Qeltrix
- Test Results: https://github.com/Qeltrix/test-poc-1
Copyright & License
© 2025 @hejhdiss (Muhammed Shafin P)
test_qeltrix.py and associated test files in the test-poc-1 repository are licensed under the GNU General Public License v3.0 (GPLv3). See the repository for full license details.
This is a community-driven project. The author provides no guarantee of updates from himself. All progress depends on community contributions.
Top comments (0)