NEXUS-Multi-Agent-RAG-on-GPU EbiAraz · PARTIAL
NEXUS is a modular Retrieval-Augmented Generation studio. Five specialist agents run in sequence, retrieval is dense (CPU MiniLM so VRAM stays with the generator), and answers stream from a **vLLM** engine on your GPU. This project lives next to the existing `vllm-env` virtualenv and is sized for a laptop GPU such as the **RTX 4050 6 GB**.
github.com/EbiAraz/NEXUS-Multi-Agent-RAG-on-GPU · ★ 1 · Forks 0 · Size 36 KB
SUMMARY
Technologies 9
Scored 8
Observed 8
Practices 6
Evidence 10
Skips 1
COVERAGE
Analyzed 47 files · 1 commits · 0 API calls
TECHNOLOGIES & DEPTH
Shell LANGUAGE Depth 70
1 files · PRODUCTION
Markdown LANGUAGE Depth 70
5 files · PRODUCTION
JavaScript LANGUAGE Depth 70
7 files · PRODUCTION
Python LANGUAGE Depth 70
26 files · PRODUCTION
pip BUILD_TOOL Depth 80
1 files · CONFIGURATION
FastAPI FRAMEWORK Depth 62
2 files · PRODUCTION
NumPy LIBRARY Depth 62
2 files · PRODUCTION
Pydantic LIBRARY Depth 39
1 files · PRODUCTION
python-dotenv LIBRARY Depth —
0 files · config only
PRACTICES
documentation · observedautomated_tests · absentcontinuous_integration · absentcontainerization · absentlinting · absentformatting · absent
ACTIVITY & OWNERSHIP
First commit 2026-09-12
Last commit 2026-09-12
Active months 1
Commits 1