brilliant13/BrainTrace_English

Python

0

230 commits

updated Jun 13, 2025

See the code

README

🎥 Demo Video

Brain-Trace Demo Video

Image

🧠 Brain‑Trace

On‑device, knowledge‑graph‑based AI system
Handling data more securely, accurately & intelligently


📌 Overview

Brain‑Trace is an on‑device AI solution that automatically analyzes documents—PDFs, audio, text, etc.—to build a knowledge graph and provides contextual question answering.
It processes everything locally, without external data transfer, and delivers source‑based answers for reliability and transparency.


🌟 Key Features

🔐 On‑Device AI: More Secure, More Cost‑Effective

  • Fully local processing with no external data transfer
  • Privacy assured at no extra cost

📖 Source‑Based Answers

  • Returns key nodes (sources) used in answers to ensure trust and transparency

🌐 Precision Inference with Graph‑RAG

  • Find nodes relevant to the query → trace logical paths through relationships → generate structured answers

📂 Automatic Document Upload & Graph Creation

  • Upload PDFs, audio, or text → auto‑analysis → visualized & searchable knowledge graph
  • Enter keywords or concepts → quickly retrieve relevant nodes and document info via similarity search

📊 RAG vs Graph RAG

MethodPrecision (P)Recall (R)F1 Score
Community‑GraphRAG (Local)67.264.8964.6
RAG66.3463.9963.88

🏆 Community‑GraphRAG (Local) achieves the best performance in multi‑hop question answering, demonstrating accurate multi‑hop reasoning.
Source: arXiv:2502.11371


🛠️ Tech Stack

ComponentTechnology
LanguagesPython, JavaScript
FrameworksFastAPI, React, Electron
DatabasesNeo4j, Qdrant, SQLite
AI ModelsKoE5, LLaMA, Whisper
ToolsVS Code

brilliant13/BrainTrace_English

Python

0

230 commits

updated Jun 13, 2025

See the code

README

🎥 Demo Video

Brain-Trace Demo Video

Image

🧠 Brain‑Trace

On‑device, knowledge‑graph‑based AI system
Handling data more securely, accurately & intelligently


📌 Overview

Brain‑Trace is an on‑device AI solution that automatically analyzes documents—PDFs, audio, text, etc.—to build a knowledge graph and provides contextual question answering.
It processes everything locally, without external data transfer, and delivers source‑based answers for reliability and transparency.


🌟 Key Features

🔐 On‑Device AI: More Secure, More Cost‑Effective

  • Fully local processing with no external data transfer
  • Privacy assured at no extra cost

📖 Source‑Based Answers

  • Returns key nodes (sources) used in answers to ensure trust and transparency

🌐 Precision Inference with Graph‑RAG

  • Find nodes relevant to the query → trace logical paths through relationships → generate structured answers

📂 Automatic Document Upload & Graph Creation

  • Upload PDFs, audio, or text → auto‑analysis → visualized & searchable knowledge graph
  • Enter keywords or concepts → quickly retrieve relevant nodes and document info via similarity search

📊 RAG vs Graph RAG

MethodPrecision (P)Recall (R)F1 Score
Community‑GraphRAG (Local)67.264.8964.6
RAG66.3463.9963.88

🏆 Community‑GraphRAG (Local) achieves the best performance in multi‑hop question answering, demonstrating accurate multi‑hop reasoning.
Source: arXiv:2502.11371


🛠️ Tech Stack

ComponentTechnology
LanguagesPython, JavaScript
FrameworksFastAPI, React, Electron
DatabasesNeo4j, Qdrant, SQLite
AI ModelsKoE5, LLaMA, Whisper
ToolsVS Code