AutoFlow - AI Ops Manager for MSMEs

An AI-powered automation builder that turns natural language into fully functional business workflows for MSMEs.

Description

AutoFlow is an open-source AI Ops automation engine designed for MSMEs that converts plain natural language instructions into deployable business workflows.

Using a multi-phase AI pipeline, AutoFlow builds end-to-end automation graphs-including triggers, conditions, AI reasoning nodes, Google Sheets lookups, WhatsApp actions, and error handling-without requiring any technical knowledge.

This enables non-technical users such as shop owners, clinics, restaurants, and service providers to automate daily tasks across WhatsApp, Instagram, Email, Web, and internal systems within seconds.


✨ Key Features

  • AI Generates Complete Automation from Text
    Users type needs like “Create an agent that accepts orders and checks stock” - AutoFlow builds the workflow graph automatically.

  • Visual Drag-and-Drop Workflow Editor
    Clean, editable workflow editor for customizing nodes.

  • Inventory-Aware Agents (Unique USP)
    Automatically connects to Google Sheets for inventory, orders, logs, etc.

  • Instant Deployment to WhatsApp
    One-click QR deployment to go live instantly.

  • Auto-Explainer Feature
    AI explains every node and action in simple language for transparency.

  • Multi-Channel Support
    WhatsApp, Instagram, Webhooks, Google APIs, Email, PayPal, Salesforce, etc.

Target Users

  • MSME owners

  • Non-technical staff

  • Solo entrepreneurs

  • Local service providers

  • Fast-growing businesses (50–500 daily orders)

Problem AutoFlow Solves

  • Too many disjoint tools

  • Automation platforms require coding knowledge

  • Custom AI agents take weeks to build

  • No WhatsApp-first solutions exist

  • Manual setup takes hours
    AutoFlow solves all these by automating everything from text.

How AutoFlow Works (3-Phase Architecture)

Phase 1 - Understanding User Intent

User describes the workflow in natural language.

Phase 2 - AI Builds Workflow Graph

Generates:

  • Triggers

  • AI reasoning nodes

  • Conditions

  • Sheet lookups

  • WhatsApp actions

  • Error handling

Phase 3 - One - Click Deployment

User scans QR → Automation goes live.


⚙ Tech Stack

  • AI Layer: Gemini 2.0 Flash, Pro, Flash Lite

  • Compute: Cloud Run, Cloud Functions

  • Database: Firestore, Realtime DB

  • Authentication: Firebase Auth

  • APIs: Google Sheets, Cloud Vision, Pub/Sub

  • Monitoring: Cloud Logging

Issues & Pull Requests Thread
No issues or pull requests added.