Web App
Sentio+ is an AI-powered decision-support platform that transforms large-scale, unstructured customer review data into actionable business insights using a Retrieval-Augmented Generation (RAG) architecture. Designed as an internal intelligence tool for Product, CX, Strategy, and Leadership teams, it enables aspect-level reasoning over customer feedback, grounding every insight in real review evidence.
Walkthrough
Thousands of reviews, reduced to what actually matters.
How it works
Fifty thousand reviews, one clear picture.
Analytics
See how ratings move, and why
Sentio+ turns 50,000 Google Play reviews into a dashboard of review volume, the rating breakdown, the sentiment split and average rating over time. Low-star reviews are grouped into complaint themes, so the biggest problem is the first thing you see.
Ask Sentio+
Ask your reviews a question in plain English
A LangGraph agent searches the review collection with its own tools and answers with real counts and a quoted review. It also says what to fix first, and it remembers the conversation for follow-ups.
Live view
Your customers and reviews on one globe
Upload a customer CSV and every row lands on an interactive globe beside the reviews mapped by country. The side panel shows where customers are and how each region rates the app.
Purpose
Serves as a centralized "Customer Intelligence Platform" for organizations managing large volumes of customer feedback. Unlike systems that rely solely on manual tagging and keyword search, Sentio+ uses RAG to enable semantic search and natural language queries across all reviews. It helps teams navigate complex feedback portfolios, track sentiment trends, and extract insights from historical reviews.
Key Features
AI-Powered Semantic Search
Semantic search across all reviews using vector embeddings, enabling natural language queries that understand context and meaning.
Conversational Q&A
Chat interface powered by LangGraph agents that can answer questions about reviews with citations to source documents.
Automated Document Processing
Text extraction and cleaning from CSV datasets, intelligent chunking for optimal retrieval, and metadata enrichment.
Aspect-Level Sentiment
Identifies specific product aspects driving sentiment, tracks trends across time and categories with evidence-grounded insights.
Metadata-Aware Retrieval
Filter by app name, category, rating, date range with LLM-powered source selection for focused queries.
Vector Store Flexibility
Persistent local ChromaDB for development, HTTP client for distributed deployments, and ChromaDB Cloud for production.
Business Intelligence
Collection statistics, trend detection across time periods, category and rating-based analysis, and export-ready insights.
Multi-Interface Access
Next.js web application for production, Streamlit demo for testing, and RESTful API for programmatic access.
Tech Stack