Skip to content
    Back to all Bounties

    Earn 13,500 ($135.00)

    Time Remainingdue 2 years ago
    Completed

    Build a Replit for a Multi-User Chatbot with Langchain, Pinecone, and Next.JS

    Sankaranaraya18
    Sankaranaraya18
    Posted 2 years ago
    This Bounty has been completed!
    @Sankaranaraya18's review of @boltgaming
    5.0
    Average Rating
    Communication 5/5, Quality 5/5, Timeliness 5/5
    “I'm really happy with John's work. Will definitely recommend him for future work.”

    Bounty Description

    Hello, verified bounty hunters!

    I'm looking to hire someone to create a Replit based on this GitHub repository (https://github.com/pinecone-io/chatbot-demo) for a multi-user chatbot built with Langchain, Pinecone, and Next.JS. The goal is to have a chatbot that runs on Replit where I can upload my own data (preferably through an UI, if not that's fine) and chat with the chatbot.

    Project Details:

    The original tutorial (https://www.pinecone.io/learn/javascript-chatbot/) outlines how to build a multi-user chatbot using Langchain, Pinecone, and Next.JS, providing a natural conversation experience with multiple users, utilizing user-based conversation history for context. The architecture consists of the chatbot, the indexer, and the Pinecone index.

    Key components:

    1. Prompt generation using PromptTemplate and LLMChain provided by Langchain.
    2. Indexer crawls a content source and deals with rate limits.
    3. Inquiry building, document retrieval, summarization (optional), and final answer construction for the chatbot.
    4. Ably integration for handling real-time communication and messaging (ably might not work on replit, if not, please figure out a different approach like socket.io, clientJS, and postgreSQL)

    Deliverables:

    1. A Replit based on the GitHub repo linked above, with necessary modifications to make it compatible with Replit.
    2. Instructions on how to upload custom data to the pinecone through replit.
    3. A demo of the chatbot working on the Replit platform, showcasing the multi-user functionality.
    4. Documentation detailing the modifications made, if any, to the original codebase, and how the Replit interacts with the chatbot, indexer, and Pinecone index.

    I'm excited to see your proposals and work with one of you to bring this multi-user chatbot to life on the Replit platform!

    P.S., If you feel the bounty amount is not adequate, please let me know. We will figure something out.