Front page, create workflow automations with prompt
Creating dynamic knowledge base for RAG
chose other workflows from sidebar
One click authentication
prompt will get refined and visualize the workflow
edit tool configurations by clicking on nodes

SIGMOYD: Natural Language-Driven AI Workflow Automation

Inspiration

Problem
Modern professionals often struggle with repetitive workflows and unstructured communication—especially in customer support across vast and dynamic data sources. Traditional automation requires technical know-how, limiting accessibility.
There is a growing need for a natural language-based platform that allows effortless task automation.

What It Does

Solution
SIGMOYD is a prompt-driven platform that enables users to create, visualize, and deploy intelligent agentic workflows. These AI agents integrate with tools like:

Gmail
Notion
GitHub
Slack
RAG Tools
External APIs
Web platforms
Custom MCP Servers

Users can trigger actions via events or prompts, participate in community chats, and maintain transparency, memory, and secure cross-platform automation.

Target Audience

SIGMOYD is designed for teams and individuals looking for smarter work execution:

Product Managers
Operations Leads
Customer Support Teams
Sales & Marketing Analysts
IT Teams
Solo Founders
Government Agencies

The platform offers no-code automation for non-technical users and developer-friendly customizations for technical teams.
Perfect fit for startups, fintech, web3, logistics, and remote-first organizations.

How We Built It

Used async capabilities of fastapi (python)
React frontend
Redis, celery for background processing, workflow log display using socket.
Clerk for auth
Composio for mcp server and tool integration
Deepseek. Gemini, groq for workflow creation
Dynamodb database
Gcp for gmail trigger (pub/sub)
Aws for deployment
Ngrok for local deployment to recieve new mails.

Challenges We Ran Into

Designing optimal prompt structures for diverse workflows
Visualizing complex workflows in an intuitive UI
Handling errors gracefully during workflow execution
Prioritizing feature integration for MVP

Accomplishments That We're Proud Of

Successfully visualized and executed workflows within 60 seconds—from creation to deployment
Won 1st prize in the Industry Academia Meet for this project

What We Learned

We deepened our understanding of:

Background task execution
Redis Pub/Sub architecture
Message queues
Prompt engineering
Real-time collaborative coding (vibe coding)

What's Next for SIGMOYD

Custom Community Chat Interface
Chat agents will trigger workflows and respond with relevant context.
Example: A CEO asks about a lead's status → the agent fetches data from HubSpot or Notion and replies intelligently.
Dynamic Knowledge Base Integration
Users can create workflows that retrieve the most relevant answers from long unstructured documents, collections, or URLs.
Triggers can include:
- Slack messages
- Incoming customer emails via Gmail
- Webhooks and other real-time events