Development of Aphra AI-Powered Workspace

Project overview : Aphra is an AI-powered research and writing workspace that turns messy notes into publish-ready content. Import docs, links, or PDFs; Aphra summarizes, cites sources, and suggests outlines in real time. Collaborate with comments and version history, then export to CMS or PDF. Privacy-first—data stays encrypted in your cloud. Ideal for product docs, blogs, and knowledge bases.

The main goal of APHRA was to streamline research-to-writing. It turns messy notes, docs, and links into clear outlines and draft content with accurate citations. By automating summarization and structure while keeping data private APHRA helps teams produce publish-ready docs faster and with less friction.

We built Aphra as a RAG pipeline: ingest docs/URLs/PDFs, normalize, chunk, and embed into a vector store; retrieval feeds LLM prompts for summaries, citations, and outline drafts. A Next.js frontend streams results via WebSockets; a Node/FastAPI backend orchestrates workers and queues. Data stays tenant-isolated with KMS-backed encryption, role-based access, and audit logs. Continuous evals and tracing tune prompts and ranking.

Features

NextJS-in-Electron App

Windows + macOS Support

Auto-update Support via AWS S3

Windows Code Signing via AWS KMS

macOS Code Signing + Notarizing

Github Actions CI/CD

Internet Speed Monitoring Support

Skills: Electron, React, TypeScript, Webpack, AWS S3, AWS KMS, @electron/notarize, electron-builder, electron-store, electron-updater