NLPAI★ FeaturedPublic repo

Knowledge Graph-based Chatbot

Transforms unstructured text (specifically Morocco's ministerial decree on digitizing public procurement) into a queryable knowledge graph. Llama 3.1 (8B, served via Ollama) extracts entities and relationships from the source text, which are incrementally stored as triplets in Neo4j. User questions in natural language are converted to Cypher queries via Langchain, retrieving relevant graph relationships (full-text indexed) and generating grounded answers with source attribution, reducing hallucination compared to a plain LLM. Built and run on Google Colab.

Technologies

PythonLLaMA 3.1LangchainNeo4jRAGGradio

Gallery

Links

Code

More like this

★

Sales To User Data Q&A Assistant Agent

Local-LLM chat assistant that turns a plain-language sales question into an auditable SQL query against 3.1M real Caterpillar-dealer records with guardrails built around five verified data traps that silently return a wrong number if ignored.

PythonFastAPIDuckDBOllama (Local LLM)SQLsqlglotJavaScript,SSE

Private repoView details

★

TAE Multilingual Analyst Agent

Offline agent that turns plain French, English, or Arabic questions into verified SQL over a 117K-row Caterpillar support-claims database, validated at 87% execution accuracy on a 100-question, 3-language benchmark.

PythonStreamlitOllamaQwenSQLiteNetworkXNumPysqlglot

Private repoView details

AI Chatbot powered by Gemini

Full-stack MERN chatbot using Google's Gemini-1.5-Flash model, with authentication, chat history, and image uploads.

MongoDBExpress.jsNode.jsReactViteGemini APIClerkImageKit

Public repoCode