This project provides a Retrieval-Augmented Generation (RAG) demo for a bank's customer support bot, targeting developers building GenAI applications. It showcases how to use the Multinear platform to ensure reliability and guard against hallucinations in customer-facing AI.

How It Works

The system employs a RAG engine built with LangChain and OpenAI for document ingestion, indexing, and query processing. A FastAPI backend serves chat, reindexing, and session management endpoints, with an HTML/React frontend. The core innovation lies in the Multinear platform, which facilitates systematic experimentation and evaluation of GenAI applications by defining specific test tasks and running them against the bot.

Quick Start & Requirements

• Installation: Clone the repository, create a .env file with OPENAI_API_KEY, and set up the environment using uv sync or pyenv/venv.
• Running the App: Execute uv run main.py or python main.py. Access the demo at http://127.0.0.1:8080.
• Running Experiments: Use uv run multinear web_dev or multinear web_dev. Access the experimentation platform at http://127.0.0.1:8000.
• Prerequisites: Python 3.9+, OpenAI API key. Jupyter Notebook support available.

Highlighted Details

• Demonstrates RAG for customer support using LangChain and OpenAI.
• Integrates with the Multinear platform for evaluation and reliability testing.
• Includes a FastAPI API server and a React frontend.
• Supports LLM tracing with Arize Phoenix.

Maintenance & Community

• Developed by Multinear.
• No specific community links (Discord, Slack) or roadmap details are provided in the README.

Licensing & Compatibility

• Licensed under the MIT License.
• Permissive license suitable for commercial use and integration into closed-source projects.

Limitations & Caveats

The demo focuses on a proof-of-concept and may not represent a production-ready system. Reliability and handling of ambiguous or off-topic questions are presented as challenges addressed by the Multinear platform, implying the base demo might have limitations in these areas.