🚀 Inspiration
We noticed how hard it is for non-coders and even devs to automate tasks using tools like Zapier, Make, and n8n. Most require complicated flows, premium plans, or clunky UIs. So we built RoboRunX — a smarter, AI-powered automation engine that works from a single prompt and builds actions behind the scenes.
🛠️ What We Built
RoboRunX is an LLM-integrated backend engine that generates automation bots from natural language. It uses LangGraph for visual flows, an LLM for command parsing, and executes actions across platforms (like Twitter, Notion, Gmail, and Google Sheets). With just one click, it can post tweets, fetch data, trigger emails, or run bots — all powered by an internal component engine.
💡 What We Learned
We explored OAuth 2.0 flows with PKCE, mastered backend flows with Flask + LangGraph, and integrated LLMs to handle user input dynamically. We also learned how to handle token refresh securely and scale modular workflows.
❌ Challenges
- OAuth integration with Twitter took a lot of trial and error.
- Building the backend engine for multiple steps with metadata was tricky.
- Debugging token expiry and retweet logic was time-consuming.
- Writing generalized components that work across use cases was a major lift.
But in the end, it was worth it.
🧠 What’s Next
We’re building it for every goddam App that exist in the world till date And also we will be Building an AI Agent System that will run smart workflows continuously — just give a task and let RoboRunX work forever in the background.


Log in or sign up for Devpost to join the conversation.