Dear ResilientDB Developers, My name is Jiazhi Sun. I am writing to introduce two projects I built on top of *ResilientDB*, and to share how they explore different application-level use cases of the ResilientDB KV service. ResShare
*ResShare – A Decentralized File Sharing Application with AI Document Assistant* ResShare is a decentralized file sharing system that combines *ResilientDB* for metadata management and *IPFS* for decentralized file storage. On top of the storage layer, it integrates a *Retrieval-Augmented Generation (RAG)* based AI assistant that allows users to query their own uploaded documents. *Key ideas* - ResilientDB is used as a *trusted, decentralized metadata store* for files, folders, ownership, and sharing relationships. - IPFS is used for actual file storage, avoiding centralized storage servers. - An optional AI layer enables private, per-user document Q&A without leaking data across users. *Core features* - User authentication and account management - File upload/download, folder management, and user-to-user sharing - Decentralized storage using IPFS, with metadata stored in ResilientDB - AI-powered document Q&A (PDF / DOCX / TXT), with source attribution and per-user isolation *Architecture highlights* - Metadata layer: ResilientDB KV service - Storage layer: IPFS - Backend: Python Flask - Frontend: React + Material UI - AI stack: Sentence Transformers, FAISS, LangChain, optional Gemini or local LLMs Repository: https://github.com/NoBugInMyCode/ResShareDeployable ------------------------------ ResChat *ResChat – A Fully Decentralized and Encrypted Chat System on ResilientDB* ResChat is a decentralized chat system built directly on the *ResilientDB blockchain*, without relying on any centralized server. Messages are transmitted and stored on-chain using the KV service, and end-to-end security is guaranteed through *RSA + AES hybrid encryption*. *Key ideas* - No central server: all messages are written to and retrieved from ResilientDB. - Privacy-first: no plaintext messages or personal data are stored anywhere. - Flexible deployment: can run on a private ResilientDB network (intranet) or a public one. *Design highlights* - Messages are stored in a custom Page structure (20 messages per page). - Page keys are deterministically derived from sender/receiver usernames to ensure both sides access the same message stream. - Messages and files are encrypted with AES-128; AES keys are encrypted separately with sender’s and receiver’s RSA public keys. - Only sender and receiver can decrypt messages or files; keys are never exposed. *Tech stack* - ResilientDB KV service - RSA + AES encryption - Python backend with Bazel-built pybind KV bindings - React frontend Repository: https://github.com/ResilientApp/ResChat ------------------------------ Motivation These two projects explore *how ResilientDB can serve as a general-purpose decentralized application backbone*, supporting: - secure metadata storage, - decentralized communication, - privacy-preserving file sharing, - and higher-level applications such as AI-assisted document interaction. I hope these projects can be useful references or discussion points for the community, especially for developers interested in application-layer designs on top of ResilientDB. Thank you for your time, and I would be happy to answer any questions or receive feedback. Best regards, Jiazhi Sun
