{"id":49663494,"url":"https://github.com/r4stin/kg-research-agent","last_synced_at":"2026-05-06T14:40:38.263Z","repository":{"id":325075368,"uuid":"1095989223","full_name":"r4stin/kg-research-agent","owner":"r4stin","description":"Evidence-grounded, multi-agent research assistant that performs RAG over scientific papers, extracts structured claims, builds a Neo4j knowledge graph, and answers questions with verifiable citations and stateful session memory.","archived":false,"fork":false,"pushed_at":"2025-11-26T21:04:02.000Z","size":60,"stargazers_count":0,"open_issues_count":0,"forks_count":0,"subscribers_count":0,"default_branch":"main","last_synced_at":"2025-11-29T16:41:51.866Z","etag":null,"topics":["adk","ai-agents","chroma","evidence-extraction","google-gemini","knowledge-graph","llm","multi-agent-system","neo4j","nlp","rag","research-tools"],"latest_commit_sha":null,"homepage":"","language":"Python","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":"mit","status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/r4stin.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":null,"funding":null,"license":"LICENSE","code_of_conduct":null,"threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":null,"support":null,"governance":null,"roadmap":null,"authors":null,"dei":null,"publiccode":null,"codemeta":null,"zenodo":null,"notice":null,"maintainers":null,"copyright":null,"agents":null,"dco":null,"cla":null}},"created_at":"2025-11-13T19:38:27.000Z","updated_at":"2025-11-26T21:04:05.000Z","dependencies_parsed_at":null,"dependency_job_id":null,"html_url":"https://github.com/r4stin/kg-research-agent","commit_stats":null,"previous_names":["r4stin/kg-research-agent"],"tags_count":5,"template":false,"template_full_name":null,"purl":"pkg:github/r4stin/kg-research-agent","repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/r4stin%2Fkg-research-agent","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/r4stin%2Fkg-research-agent/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/r4stin%2Fkg-research-agent/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/r4stin%2Fkg-research-agent/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/r4stin","download_url":"https://codeload.github.com/r4stin/kg-research-agent/tar.gz/refs/heads/main","sbom_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/r4stin%2Fkg-research-agent/sbom","scorecard":null,"host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":286080680,"owners_count":32699109,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2026-05-06T08:33:17.875Z","status":"ssl_error","status_checked_at":"2026-05-06T08:33:17.221Z","response_time":117,"last_error":"SSL_connect returned=1 errno=0 peeraddr=140.82.121.6:443 state=error: unexpected eof while reading","robots_txt_status":"success","robots_txt_updated_at":"2025-07-24T06:49:26.215Z","robots_txt_url":"https://github.com/robots.txt","online":false,"can_crawl_api":true,"host_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub","repositories_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories","repository_names_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repository_names","owners_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners"}},"keywords":["adk","ai-agents","chroma","evidence-extraction","google-gemini","knowledge-graph","llm","multi-agent-system","neo4j","nlp","rag","research-tools"],"created_at":"2026-05-06T14:40:36.127Z","updated_at":"2026-05-06T14:40:38.251Z","avatar_url":"https://github.com/r4stin.png","language":"Python","funding_links":[],"categories":[],"sub_categories":[],"readme":"\n# 📚 **KG-Research-Agent**\n### *Multi-Agent, Evidence-Grounded Research System with Gemini, ADK, ChromaDB \u0026 Neo4j*\n\n\u003cdiv align=\"center\"\u003e\n\n**🔥 A research-grade AI agent that extracts claims + evidence from scientific papers, stores them in a knowledge graph, retrieves context, and answers questions using multi-agent reasoning with session memory.**\n\n[![Python](https://img.shields.io/badge/Python-3.10-blue.svg)]()\n[![License: MIT](https://img.shields.io/badge/License-MIT-green.svg)]()\n[![Neo4j](https://img.shields.io/badge/Neo4j-GraphDB-blue.svg)]()\n[![ChromaDB](https://img.shields.io/badge/ChromaDB-Vector_Store-purple.svg)]()\n[![Gemini](https://img.shields.io/badge/Gemini-LLM-orange.svg)]()\n\n\u003c/div\u003e\n\n---\n\n# 🚀 **Overview**\n\n**KG-Research-Agent** is an AI-powered research assistant that:\n\n- Ingests scientific PDFs  \n- Embeds + stores them in ChromaDB  \n- Retrieves relevant text chunks (RAG)  \n- Extracts **structured claims \u0026 evidence** from papers  \n- Stores them in a **Neo4j Knowledge Graph**  \n- Answers questions using **citations grounded in source text**  \n- Uses a **multi-agent pipeline** (Planner → Retriever → Evidence → Answer)  \n- Supports **multi-turn conversations with session memory**\n\nA full walkthrough of the multi-agent research system is available on YouTube:\n\n👉 **[Watch the Concept Overview](https://youtu.be/vaq0-AMOudo)**\n\n---\n\n# 🧠 **Updated Architecture (Multi-Agent + Memory)**\n\n```\n┌──────── User ────────┐\n          │\n          ▼\n┌───────────────┐\n│ Planner Agent │  ← uses chat history + memory\n└───────────────┘\n     │ plans tasks\n     ▼\n┌────────────────────────┐\n│ Retriever Agent        │ → ChromaDB (vector search)\n└────────────────────────┘\n     │ chunks\n     ▼\n┌────────────────────────┐\n│ Evidence Agent         │ → extracts claims + sentences\n└────────────────────────┘\n     │ structured JSON\n     ▼\n┌────────────────────────┐\n│ Answer Agent           │ → composes human-readable answer\n└────────────────────────┘\n     │\n     ▼\n **Final Answer + Citations**\n\n📦 Persistent Storage:\n- Neo4j → long-term knowledge graph\n- ChromaDB → vector retrieval\n- SessionState → short-term conversation memory\n```\n\n---\n\n# ✨ **Current Features**\n\n### ✔️ PDF → Chunking → Vector Storage  \n### ✔️ RAG Retrieval (Chroma + Gemini)  \n### ✔️ Multi-Agent System (Planner → Retriever → Evidence → Answer)  \n### ✔️ Structured JSON Evidence Extraction  \n### ✔️ Neo4j Knowledge Graph Storage  \n### ✔️ Session Memory (short-term conversational context)  \n### ✔️ Deduplication (per chunk + semantic similarity)  \n### ✔️ Multi-turn conversational research workflow  \n\n---\n\n# 🏁 **Getting Started**\n\n## 1️⃣ Clone the Repo\n```\ngit clone https://github.com/yourusername/kg-research-agent.git\ncd kg-research-agent\n```\n\n## 2️⃣ Create Conda Environment\n```\nconda create -n kg-research-agent python=3.10\nconda activate kg-research-agent\n```\n\n## 3️⃣ Install Requirements\n```\npip install -r requirements.txt\n```\n\n## 4️⃣ Environment Variables (`.env`)\n\n```\nGOOGLE_API_KEY=\"your-key\"\nCHROMA_DB_PATH=\"data/chroma\"\nPDF_STORAGE=\"data/papers\"\n\nNEO4J_URI=\"bolt://localhost:7687\"\nNEO4J_USER=\"neo4j\"\nNEO4J_PASSWORD=\"yourpassword\"\n```\n\n---\n\n# 🧪 **Running the System**\n\n### PDF Ingestion\n```\npython -m src.tools.pdf_ingest\n```\n\n### Evidence Extraction\n```\npython -m src.run_evidence_extraction\n```\n\n### KG Query\n```\npython -m src.pipelines.run_kg_query\n```\n\n# 🔧 **New: Multi-Agent Runner**\n\nRun full pipeline with memory:\n\n```\npython -m src.pipelines.run_multi_agent_pipeline\n```\n\nExample:\n\n```\nYou: What is a major challenge in scholarly information retrieval?\nYou: Summarize in one sentence.\n```\n\nThe agent maintains context across turns.\n\n---\n\n# 🗺️ **Roadmap**\n\n## 🟥 Agent Quality (Next Milestone)\n- ADK logs + traces\n- Metrics for agent performance\n- LLM-as-a-Judge evaluation\n\n## 🟦 Multi-Agent Enhancements\n- Add **KG Agent** (read/write Neo4j in pipeline)\n- Add planner task types: `kg_query`, `kg_write`\n- Context compaction + memory optimization\n\n## 🟩 Productionization\n- A2A protocol (agent-to-agent messaging)\n- Deployment to **Vertex AI Agent Engine**\n- API endpoints + orchestration layer\n\n---\n\n# 📜 License\n\nMIT License.  \nYou may use, modify, and distribute this project freely.\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fr4stin%2Fkg-research-agent","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fr4stin%2Fkg-research-agent","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fr4stin%2Fkg-research-agent/lists"}