Skip to content
/ rag_bot Public

A specialized RAG bot using Cohere/OpenAI/Pinecone/CrewAI/Gladia

Notifications You must be signed in to change notification settings

hvbr1s/rag_bot

Repository files navigation

Demo

Watch the video

Demo (with voice recognition)

Watch the video

Intro

SamBot is an AI-powered customer service bot that utilizes Retrieval-Augmented Generation (RAG) to accurately troubleshoot user issues. It is designed for high reliability and low maintenance, with the capability to handle traffic asynchronously.

How to run the bot locally for testing purposes

  1. Make sure you have the correct .env variables: OPENAI_API_KEY, BACKEND_API_KEY, API_KEY_NAME, PINECONE_API_KEY, PINECONE_ENVIRONMENT, COHERE_API_KEY.
  2. From the root folder, run uvicorn app:app --reload --port 8800 to start a local instance of the bot
  3. Using Postman ping the http://127.0.0.1:8800/gpt endpoint with a request formatted as follow:
POST /gpt HTTP/1.1
Host: <your_url>
Authorization: Bearer <BACKEND_API_KEY>
Content-Type: application/json

{
    "user_input": "your_question",
    "user_locale": "your_locale",
    "user_id": "any_number"
}
  1. After a brief moment (10-20 seconds), you should see the bot's response to your question appear in the console. You can test different locales by changing user_locale to eng, fr, etc.

About

A specialized RAG bot using Cohere/OpenAI/Pinecone/CrewAI/Gladia

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published