Aug 19, 2023
AI-powered Slackbot for task management in Slack, saving 30+ mins/day, used by 200+ people.
Overview
TwingBot is an intelligent Slack bot that revolutionizes task management within teams. By leveraging AI capabilities, it helps users manage their tasks, schedules, and workflows directly within Slack, eliminating the need to switch between multiple productivity tools.
Key Features
- 🤖 AI-Powered Task Creation - Natural language processing for task creation
- ⏰ Smart Scheduling - Automatic task prioritization and scheduling
- 👥 Team Collaboration - Seamless task assignment and tracking
- 📊 Analytics Dashboard - Productivity insights and reporting
- 🔔 Smart Notifications - Context-aware reminders and updates
Impact
- 30+ minutes saved per day per user
- 200+ active users across multiple organizations
- Increased team productivity by reducing context switching
- Improved task completion rates through smart reminders
Tech Stack
- Backend: Python, FastAPI
- AI/ML: OpenAI GPT, LangChain
- Database: Supabase, PostgreSQL
- Task Processing: Hatchet (async task system)
- Integration: Slack API
Demo Video
Links
Part of the Twing ecosystem of productivity tools.
Jun 1, 2024
RAG-based legal research & case management platform using Azure Vector DB, OpenAI, and Azure Search.
Overview
Legaling is an advanced legal research and case management platform that leverages Retrieval-Augmented Generation (RAG) technology to help legal professionals find relevant cases, statutes, and legal precedents quickly and accurately.
Key Features
- 🔍 Intelligent Legal Search - RAG-powered search across legal databases
- 📚 Case Management - Comprehensive case tracking and documentation
- 🤖 AI Legal Assistant - Natural language queries for legal research
- 📊 Citation Analysis - Automatic citation parsing and validation
- 🔒 Secure Document Handling - Enterprise-grade security for sensitive legal documents
Technical Architecture
- AI Framework: RAG (Retrieval-Augmented Generation)
- Vector Database: Azure Vector DB for semantic search
- AI Model: OpenAI GPT for natural language processing
- Search Engine: Azure Cognitive Search
- Backend: Python, FastAPI
- Database: PostgreSQL with vector extensions
Use Cases
- Legal Research: Find relevant cases and precedents
- Case Management: Track case progress and documents
- Document Analysis: Extract key information from legal documents
- Citation Validation: Verify and format legal citations
Demo Video
Links
Revolutionizing legal research with AI technology.
Oct 8, 2024
AI-powered system for template-based data extraction from diverse documents using OCR, language identification, and parsing algorithms.
Overview
TwingParse is an intelligent document processing system that can extract structured data from various document types using advanced AI techniques. It combines OCR technology with language identification and various parsing algorithms to handle diverse document formats.
Key Features
- 📄 Multi-Format Support - PDF, images, scanned documents, digital files
- 🌐 Language Detection - Automatic language identification and processing
- 🤖 Template-Based Extraction - Custom templates for different document types
- 👁️ OCR Integration - Advanced optical character recognition
- 🎯 High Accuracy - AI-powered parsing with validation algorithms
Technical Components
- OCR Engine: Advanced optical character recognition
- Language Processing: Multi-language text identification using
langdetect - Document Processing: Template-based parsing algorithms
- AI Models: Custom trained models for document understanding
- Backend: Python, FastAPI
- Storage: Cloud-based document storage and processing
Use Cases
- Invoice Processing - Extract key information from invoices
- Form Data Extraction - Digitize form submissions
- Document Digitization - Convert physical documents to structured data
- Compliance Documentation - Process regulatory documents
Demo Video
Links
Transforming unstructured documents into actionable data.
Oct 12, 2024
AI translation system for Indian languages that handles both scanned and digital documents via textract, tesseract, and langdetect.
Overview
TwingTranslate is a specialized AI-powered translation system designed specifically for Indian languages. It can process both scanned and digital documents, making it accessible for a wide range of use cases from government documents to business communications.
Key Features
- 🌐 Indian Language Support - Specialized for Indian regional languages
- 📄 Document Types - Both scanned and digital document processing
- 🤖 AI-Powered Translation - Context-aware translation algorithms
- 👁️ OCR Integration - Extract text from scanned documents
- 🎯 Language Detection - Automatic source language identification
Technical Architecture
- Text Extraction: Textract for digital documents
- OCR Processing: Tesseract for scanned documents
- Language Detection: langdetect for automatic identification
- Translation Engine: Various AI models for Indian languages
- Backend: Python, FastAPI
- Processing Pipeline: Async document processing
Supported Languages
- Hindi - Most widely used Indian language
- Regional Languages - Support for major Indian regional languages
- Multi-script Support - Devanagari, Tamil, Telugu, and more
- Bidirectional Translation - English ↔ Indian languages
Use Cases
- Government Documents - Translate official documents
- Business Communications - Cross-language business correspondence
- Educational Content - Academic material translation
- Legal Documents - Legal text translation with context preservation
Links
Breaking language barriers with AI-powered translation for Indian languages.
Oct 22, 2022
A custom version of a CLI zapier, imagining how internal zapier platform might work.
Overview
Custom Zapier CLI is a Python-based automation platform that enables seamless integration between various services and applications. Built as a command-line interface, it demonstrates how internal automation platforms might function, providing a foundation for building custom workflow automations.
Key Features
- 🔗 Service Integration - Connect multiple services like Google Apps, Dropbox, and email systems
- ⚡ Async Processing - Built with Celery for handling background tasks and workflows
- 🗄️ Database Management - PostgreSQL integration for reliable data storage
- 📊 Google Sheets Integration - Automated data synchronization with Google Sheets
- 📁 Dropbox Automation - File monitoring and automated downloads based on change detection
- 🔄 Workflow Engine - Core engine for managing and executing automation workflows
Technical Architecture
- Backend Framework: Python with Django
- Task Queue: Celery for async processing
- Database: PostgreSQL for data persistence
- Service APIs: Google Apps API, Dropbox API
- Architecture: Modular design with separate components for each service
Use Cases
Primary Automation Workflows
1.Email to Google Sheets
- Automatically extract data from emails
- Create new rows in Google Sheets with email data
- Perfect for lead tracking and data collection
2.Dropbox File Monitoring
- Monitor Dropbox folders for file changes
- Download files that have been modified within specified time duration
- Automated file synchronization and backup
Links
Demonstrating how internal automation platforms can be architected for scalable workflow management.