library_name: lexrag pipeline_tag: text-generation tags:
- legal-rag
- legal-ai
- legal-research
- retrieval-augmented-generation
- uae-law
- indian-law
- gst
- vat
- corporate-tax
- hybrid-search
- fastembed
- qdrant
- legal-tech
- compliance
- tax-research
- case-law
- evolucentai
- eulogik license: agpl-3.0 language:
- en datasets:
- Indian Kanoon
- UAE Legislation Portal metrics:
- retrieval-recall
- grounding-accuracy widget:
- text: "What is the GST rate on online gaming contest prizes in India?"
- text: "UAE VAT on financial services"
- text: "TDS under Section 194B on winnings"
- text: "UAE Corporate Tax free zone treatment" co2_eq_emissions: 0
⚖️ LexRAG — The Legal Intelligence Terminal
Enterprise-Grade Hybrid RAG for UAE & Indian Law, Taxation & Compliance
What is LexRAG?
LexRAG is a professional-grade, high-performance Retrieval-Augmented Generation (RAG) platform purpose-built for UAE and Indian laws, taxation, accounting standards, and corporate compliance. It combines hybrid dense-sparse retrieval, neural reranking, auto-jurisdiction detection, and multi-provider LLM streaming into a zero-latency terminal interface.
⚡ Built by Evolucent AI — Premium Legal Technology Solutions
🛠️ Engineered by Eulogik — Enterprise AI & Systems Integration
Key Capabilities
| Capability | Technology |
|---|---|
| Dense Retrieval | BAAI/bge-small-en-v1.5 via fastembed |
| Sparse Retrieval | prithivida/Splade_PP_en_v1 via fastembed |
| Hybrid Fusion | Reciprocal Rank Fusion (RRF) |
| Neural Reranker | BAAI/bge-reranker-base (CrossEncoder) |
| Vector Database | Qdrant (on-disk, no Docker required) |
| LLM Providers | Groq, OpenRouter, Ollama (dynamic model catalog) |
| Jurisdiction | Auto-detect India/UAE/Both with manual override |
| Confidence Tiers | GROUNDED / PARTIAL / SYNTHESIZED |
| Streaming | SSE with heartbeat keep-alive |
| Persistence | SQLite chat history with session management |
Quick Start
# Install from PyPI
pip install lexrag
# Or clone from source
git clone https://github.com/eulogik/LexRAG.git
cd LexRAG
pip install -r requirements.txt
# Configure
cp .env.example .env
# Edit .env with your API keys
# Run
python -m uvicorn api.main:app --host 0.0.0.0 --port 8000
# Open http://localhost:8000
Architecture
flowchart TD
UI[HTML/CSS/JS SPA] <-->|SSE / API| FastAPI[FastAPI Server]
FastAPI <-->|History| SQL[(SQLite)]
Query --> Detect[Auto-Jurisdiction]
Detect --> Search{Hybrid Search}
Search -->|Dense| Dense[BGE-small-en-v1.5]
Search -->|Sparse| Sparse[Splade_PP_en_v1]
Dense & Sparse -->|RRF| Qdrant[(Qdrant)]
Qdrant -->|Candidates| Rerank[BGE-Reranker-Base]
Rerank -->|Top Docs| Context[Build Prompt]
Context --> LLM{LLM Router}
LLM -->|Cloud| Groq[Groq]
LLM -->|Cloud| OpenRouter[OpenRouter]
LLM -->|Local| Ollama[Ollama]
Ecosystem
| Platform | Link | Description |
|---|---|---|
| GitHub | eulogik/LexRAG | Source code & issues |
| PyPI | lexrag | Python package |
| HuggingFace | evolucentai/lexrag | Model card |
| Eulogik | eulogik.com | Engineering partner |
| Evolucent AI | evolucentai.com | Product & commercial |
License
AGPL v3 — Free for open-source and internal use.
Commercial licenses available from Evolucent AI for proprietary deployments.
Metadata
Release files for lexrag 3.3.0
For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.
Source distribution (sdist)
| File | Size | Uploaded | |
|---|---|---|---|
| lexrag-3.3.0.tar.gz | 25.4 kB | Details |
Built distribution (wheel)
| File | Interpreter | ABI | Platform | Reset |
|---|---|---|---|---|
| lexrag-3.3.0-py3-none-any.whl | Python 3 | none | any | Details |
Total release size: 49.5 kB
Release files / lexrag-3.3.0.tar.gz
| Download URL | lexrag-3.3.0.tar.gz |
|---|---|
| Size | 25.4 kB |
| Tags | Source |
|
SHA-256 checksum How to use checksums |
4da22808c00264f895e0cb8be6d70923924e329819ff330ffda6ddb21b83b84f
|
|
BLAKE2b-256 checksum How to use checksums |
6d7c0c88298d18f34859da0278ac590e2e268c3b43a213df648ce91235e1974d
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
twine/6.2.0 CPython/3.11.15
|
Release files / lexrag-3.3.0-py3-none-any.whl
| Download URL | lexrag-3.3.0-py3-none-any.whl |
|---|---|
| Size | 24.1 kB |
| Tags | Python 3 |
|
SHA-256 checksum How to use checksums |
bcfc6993c7df0f718b06905b078373855d77696b93c26baf80e1cde1df2cae68
|
|
BLAKE2b-256 checksum How to use checksums |
168a53c85f15dbc036a38849700aea45be56b25a2a43664fcf73eb90d0c33735
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
twine/6.2.0 CPython/3.11.15
|