# Projects | Achal Bajpai - Page: https://achal.works/projects - Markdown: https://achal.works/projects.md Featured production projects, tools, and platforms built and shipped by Achal Bajpai. ## [Peptide m/z Calculator](https://achal.works/projects/peptide-mz-calculator) - Status: live - Built: 2025 - Last shipped: 2025-08-20 - Website: https://github.com/OpenMS/mz-calculator - Repository: https://github.com/OpenMS/mz-calculator - Stack: Python, Streamlit, pyOpenMS, OpenMS, mzML, pandas, NumPy, Plotly, Docker, GitHub Actions, CI Checks, pytest A production-ready web application built for the proteomics research community. Supports mzML file processing, peptide mass-to-charge ratio calculations, and interactive data visualization. Built and deployed with OpenMS, serving thousands of researchers worldwide. Features automated CI/CD pipelines via GitHub Actions and containerized deployment with Docker. ## [FlowAgent](https://achal.works/projects/flowagent) - Status: live - Built: 2026 - Last shipped: 2026-06-12 - Repository: https://github.com/achalbajpai/flowagent - Stack: Next.js, FastAPI, LangGraph, React Flow, Rust, Redis, PostgreSQL, LiteLLM, MCP, Docker, Drizzle, Tailwind v4 A visual AI workflow builder that compiles a React Flow canvas into a real LangGraph state machine at runtime. Drag nodes onto a canvas, wire them up, and hit run: each node lights up live over WebSocket as it executes. Runs durably on a Redis-backed queue so a server restart mid-run doesn't lose work. Every published workflow is automatically registered as an MCP tool, so Claude Desktop can run it by name. Includes 14 node types: LLM (Claude via LiteLLM with fallback), web search, sandboxed Python code execution, HTTP with SSRF guard, email, condition/switch branching, human-approve with signed email links, sub-workflow, loop, merge, delay, Jinja transform, vector memory, and trigger (cron/webhook/MCP). Also ships credentials (AES-256-GCM encrypted, OAuth for Google/GitHub/Slack), team workspaces with RBAC, per-run cost tracking, monthly budget caps, analytics dashboard, and AI-powered workflow generation from plain English. ## [Symphony Music Bot](https://achal.works/projects/symphony) - Status: live - Built: 2020 - Last shipped: 2022-09-10 - Repository: https://github.com/achalbajpai/Symphony- - Stack: JavaScript, Node.js, Discord.js, FFmpeg, Lavalink, ytdl-core, YouTube API, Spotify API, SoundCloud API, MongoDB Architected and scaled from the ground up to serve 300,000+ users across 300+ servers. Features include 40+ real-time audio filters, multi-platform streaming from YouTube, Spotify, and SoundCloud, custom playlist management, and full slash command support. Maintained over 2 years with continuous performance improvements and zero-downtime deployments. ## [Terminal Portfolio](https://achal.works/projects/terminal-portfolio) - Status: live - Built: 2024 - Last shipped: 2025-03-15 - Website: https://terminal.achal.works - Stack: Next.js, React, TypeScript, Tailwind CSS, Framer Motion, App Router, Vercel, Command Palette, Keyboard Shortcuts, SEO, OG Images, Upstash Redis, Visitor Analytics A creative portfolio built to look and feel like a real terminal. Visitors navigate using command-line inputs to discover projects, experience, and contact info. Features smooth animations, keyboard shortcuts, command history, and a fully interactive shell-like interface. Built as an exploration of unconventional portfolio design. ## [Scrolla](https://achal.works/projects/scrolla) - Status: live - Built: 2025 - Last shipped: 2025-03-02 - Repository: https://github.com/achalbajpai/scrolla - Stack: TypeScript, React, Vite, Redux, Python, Flask, FastAPI, OpenAI (GPT-4o-mini, DALL·E-3), Deepgram, Crawl4AI, Playwright, FFmpeg, Firebase, Docker Scrolla turns long-form content into captivating short-form videos: endless scroll meets endless knowledge. Feed it PDFs, URLs, or web articles and it transforms them into YouTube Shorts-format video, bridging in-depth narratives and the reels-dominated feed. Generates visuals with DALL·E-3, scripts with GPT-4o-mini, and narration via Deepgram, with agentic crawling (Crawl4AI, Playwright) and FFmpeg-based media processing. Won Best with Code at SacHacks VI (UC Davis) and was runner up in Design & Code Technical Implementation. Open source on GitHub. ## [Fragments](https://achal.works/projects/fragments) - Status: live - Built: 2026 - Last shipped: 2025-05-01 - Repository: https://github.com/achalbajpai/fragments - Stack: Next.js 15, MUX, Supabase, Tailwind CSS, Shadcn UI, TypeScript Fragments is a web app and Chromium extension for creators who believe great work is shaped by intentional consumption. Ideas don't appear in isolation; they emerge from moments you notice, save, and revisit over time, and a short video clip can carry more insight than pages of text. Fragments helps you capture, organize, and rediscover those moments so they're there when you need them. Built with Mux for video processing, streaming, and AI features, and Supabase for auth and database with row-level security. Global winner of the MUX × Dev.to Worldwide Show and Tell Challenge 2025, featured by MUX and Dev.to. Open source on GitHub. ## [Wynnie](https://achal.works/projects/wynnie) - Status: live - Built: 2025 - Last shipped: 2025-08-07 - Website: https://wynnie.vercel.app - Repository: https://github.com/achalbajpai/wynnie - Stack: Next.js, AssemblyAI, OpenAI, Perplexity Sonar, SERP API, Firebase, Supabase, UPI LiteX, Tailwind CSS, PWA Wynnie is your autonomous AI shopping companion, a personal shopping agent that understands what you want through natural language and handles everything from finding products to getting the best deals. Built on AssemblyAI's speech recognition (50+ languages, real-time, word-level timestamps, speaker ID), with OpenAI for intent and a multi-agent orchestrator that researches products via Perplexity Sonar and SERP APIs, auto-applies coupons, and handles UPI LiteX payments. Google OAuth via Firebase, Supabase backend, PWA with offline support, CI/CD via GitHub Actions, and GDPR-focused security. Designed to make shopping accessible to everyone regardless of language, tech skills, or accessibility needs, especially the underserved who are shut out by interfaces built for digital natives. Global winner of the AssemblyAI Voice Agents Challenge. ## [thecontentmind](https://achal.works/projects/thecontentmind) - Status: live - Built: 2025 - Last shipped: 2025-09-15 - Repository: https://github.com/achalbajpai/thecontentmind - Stack: TanStack Start, TypeScript, PostgreSQL, Drizzle ORM, Better Auth, Nebius AI (Qwen3), OpenAI SDK, Vite, Docker, Tailwind CSS An open-source AI social-media content studio that reached 200+ users. Pick a platform, set tone, audience, and goal, hit generate, and watch three caption variants stream in live, token-by-token. Save the ones you like to a library and "more like this" from there. Three variants stream in parallel but cost a single quota unit per generate action; quotas are computed from actual rows, not a mutable counter. The AI stream filters out the model's reasoning blocks so the first real token stays sub-second. Underwent a full framework cutover from Next.js App Router to TanStack Start in stages: database, then auth, then a framework-agnostic service layer, then the framework swap, staying shippable at every step. Runs fully locally: Postgres in Docker, email/password auth, no cloud accounts required.