Add VeritasGraph to Agents → Research & Knowledge Retrieval

[bibinprathap]

awesome-datascience PR Entry

✅ Package Info

• Project: VeritasGraph
• Category: Agents → Research & Knowledge Retrieval
• GitHub: https://github.com/bibinprathap/VeritasGraph
• PyPI: https://pypi.org/project/veritasgraph-mcp/
• License: MIT
• Language: Python

---

🎯 Target Category: Agents → Research & Knowledge Retrieval

Why this fits: VeritasGraph is specifically designed for building knowledge graphs from research papers, medical literature, and policy documents. It extracts structured knowledge with verifiable source attribution - perfect for the "Research & Knowledge Retrieval" subsection.

Similar entries in this section:

• BGPT MCP - Scientific papers database for AI agents
• Chunk Tuner - Document chunking strategies for RAG
• II-Commons - Retrieval across arXiv, PubMed, policy corpora

Unique value: Only GraphRAG framework combining hierarchical tree-search with typed property graphs and 100% verifiable source attribution for research workflows.

---

📝 Entry to Add

- [VeritasGraph](https://github.com/bibinprathap/VeritasGraph) - Enterprise GraphRAG framework with MCP server for building knowledge graphs from research papers and documents. Combines tree-search retrieval with typed entity relationships and verifiable source attribution. Designed for medical research, policy analysis, and air-gapped deployments. [PyPI](https://pypi.org/project/veritasgraph-mcp/)

Insert location: In the "Agents → Research & Knowledge Retrieval" section, alphabetically (likely between "Chunk Tuner" and "II-Commons" or after "Spraay x402 Gateway")

---

📋 PR Title

Add VeritasGraph to Agents → Research & Knowledge Retrieval

---

📋 PR Description

## Summary
VeritasGraph is an enterprise GraphRAG framework for building knowledge graphs from research papers, medical literature, and policy documents with verifiable source attribution.

## Category Fit: Agents → Research & Knowledge Retrieval

VeritasGraph fits in the "Research & Knowledge Retrieval" subsection as a specialized tool for data science agents working with research literature:

### Key Features for Data Science Workflows
- **Knowledge Graph Construction**: Extracts typed entities and relationships from research papers into queryable Neo4j property graphs
- **GraphRAG Architecture**: Combines hierarchical tree-search (PageIndex) with graph-based retrieval for complex research queries
- **Verifiable Attribution**: Every answer includes exact citations to source documents (0% hallucination) - critical for scientific research
- **MCP Server**: Production-ready Model Context Protocol server (`veritasgraph-mcp` on PyPI) for AI agent integration
- **Offline Capability**: Runs entirely air-gapped via Ollama integration for secure research environments

### Use Cases in Data Science
- Medical research knowledge graphs from clinical literature
- Policy document analysis with audit-trail citations
- Scientific literature exploration with typed relationship queries
- Enterprise document retrieval for compliance and security

### Technical Details
- **Language**: Python
- **Package**: `veritasgraph-mcp` on PyPI (v0.1.2)
- **License**: MIT (Open Source)
- **Integration**: MCP server, Gradio UI, Python SDK, Docker support
- **Dependencies**: Ollama (LLMs), Neo4j (optional graphs)

## Comparison with Similar Tools

In the "Research & Knowledge Retrieval" section:
- **BGPT MCP**: Database of scientific papers from raw experimental data
- **Chunk Tuner**: Benchmarks document chunking strategies for RAG
- **II-Commons**: Retrieval across arXiv, PubMed, policy corpora
- **VeritasGraph**: GraphRAG with knowledge graphs + tree-search + typed relationships

**Differentiator**: Only framework combining hierarchical tree navigation with property graphs and 100% source attribution for research workflows.

## Quality Evidence
- **Active Development**: Recent commits (2026-08-17)
- **Published Package**: veritasgraph-mcp v0.1.2 on PyPI
- **Documentation**: Comprehensive README, architecture docs, examples
- **Production-Ready**: Docker deployment, MCP server, REST API, Gradio UI
- **Open Source**: MIT license

## Links
- **Repository**: https://github.com/bibinprathap/VeritasGraph
- **PyPI Package**: https://pypi.org/project/veritasgraph-mcp/
- **License**: MIT

---

🚀 Step-by-Step Instructions

1. Fork awesome-datascience

# On GitHub: Fork academic/awesome-datascience to bibinprathap/awesome-datascience

2. Clone and setup

cd ~/projects
git clone https://github.com/bibinprathap/awesome-datascience.git
cd awesome-datascience
git remote add upstream https://github.com/academic/awesome-datascience.git
git fetch upstream
git checkout master
git merge upstream/master

3. Create branch

git checkout -b add-veritasgraph-research

4. Edit README.md

Find the "Agents → Research & Knowledge Retrieval" section (around line 145) and add the entry alphabetically:

- [VeritasGraph](https://github.com/bibinprathap/VeritasGraph) - Enterprise GraphRAG framework with MCP server for building knowledge graphs from research papers and documents. Combines tree-search retrieval with typed entity relationships and verifiable source attribution. Designed for medical research, policy analysis, and air-gapped deployments. [PyPI](https://pypi.org/project/veritasgraph-mcp/)

5. Commit

git add README.md
git commit -m "Add VeritasGraph to Agents → Research & Knowledge Retrieval

VeritasGraph is an enterprise GraphRAG framework with MCP server for building
knowledge graphs from research papers with verifiable source attribution."

6. Push

git push origin add-veritasgraph-research

7. Create PR on GitHub

• Go to: https://github.com/bibinprathap/awesome-datascience
• Click "Compare & pull request"
• Title: Add VeritasGraph to Agents → Research & Knowledge Retrieval
• Description: Use the PR description from above
• Submit to upstream: academic/awesome-datascience

---

⚠️ Important Notes

awesome-datascience Specific Requirements

1. Format: - [Name](link) - Description with details. Optional [Resource](url)
2. Dash separator: Use regular dash (-), not bullet asterisk (*)
3. Descriptive: Include enough detail about capabilities and use cases
4. Links: Can include additional links like PyPI, docs in square brackets
5. Section fit: Must clearly fit the chosen subsection

Format Rules

• Dash with space: - [Name](link) - Description.
• Can be multiple sentences
• Can include technical details
• Can have bracketed auxiliary links at end: [PyPI](url)
• Alphabetical order within subsection (usually)

---

📊 Competitive Positioning

Similar tools in Agents → Research & Knowledge Retrieval:

1. BGPT MCP

   • Focus: Database of scientific papers from raw experimental data
   • Returns: 25+ structured fields per paper (methods, results, sample sizes)
   • Format: MCP server

2. Chunk Tuner

   • Focus: Benchmark document chunking for RAG
   • Returns: Retrieval quality scores, configuration recommendations
   • Format: Python library + MCP server

3. II-Commons

   • Focus: Deterministic retrieval across arXiv, PubMed, US policy
   • Returns: Raw documents from specific corpora
   • Format: CLI + skill

4. Spraay x402 Gateway

   • Focus: 23 research endpoints (Wikipedia, arXiv, PubMed, Wikidata)
   • Returns: Citations, entity extraction
   • Format: x402 payment gateway (pay-per-call)

VeritasGraph differentiation:

• GraphRAG: Only tool combining knowledge graphs with retrieval-augmented generation
• Tree + Graph Hybrid: Unique hierarchical PageIndex navigation + typed property graphs
• Zero Hallucination: 100% verifiable citations to source documents
• Air-Gapped: Runs entirely offline via Ollama (no external APIs)
• Enterprise Focus: Policy compliance, medical research, security-first design
• Full Stack: MCP server + UI + SDK + Docker + REST API

---

✅ Quality Checklist

• Python package (veritasgraph-mcp on PyPI)
• Active development (commits in 2026)
• Open source (MIT license)
• Comprehensive documentation
• Production-ready (Docker, MCP server, UI, API)
• Fits category (Research & Knowledge Retrieval for agents)
• Unique value (GraphRAG with source attribution)
• MCP server support (for agent integration)
• Relevant to data science workflows

---

🔗 Quick Links

• Main Repo: https://github.com/bibinprathap/VeritasGraph
• PyPI Package: https://pypi.org/project/veritasgraph-mcp/
• awesome-datascience: https://github.com/academic/awesome-datascience
• Section: Agents → Research & Knowledge Retrieval (line ~145)

---

🎯 Success Metrics

• PR created with proper format
• Entry in correct subsection
• Description highlights research/data science use cases
• PR merged

---

Last updated: 2026-08-17

[Copilot]

Pull request overview

Adds a new entry for VeritasGraph to the repo’s curated list under Agents → Research & Knowledge Retrieval, expanding the catalog of tools relevant to data-science agent workflows.

Changes:

• Add a README list item for VeritasGraph (GitHub link + PyPI link).
• Place the new entry in the “Research & Knowledge Retrieval” subsection in alphabetical order.

---

💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.

[hmert]

@bibinprathap could you follow this template please https://github.com/academic/awesome-datascience/blob/live/.github/pull_request_template.md