Inspiration
Toliq was inspired by the gap between powerful AI assistants and actual productivity tools. While AI can answer questions, most solutions lack the ability to take action in the real world. We wanted to create an AI assistant that could not only understand requests but directly interact with the digital tools people use daily, making it truly useful for busy professionals.
What it does
Toliq is an AI-powered assistant that can interact directly with Google Calendar and Google Sheets. Users can have natural conversations with Toliq and ask it to perform tasks like scheduling meetings, creating events, reading spreadsheet data, or updating information—all through a clean, intuitive chat interface. Instead of just suggesting what to do, Toliq actually does it for you.
How we built it
We built Toliq using a stack of modern technologies: Backend: Flask server with custom function calling architecture that bridges LLM capabilities with external APIs Frontend: Next.js application with a responsive, modern chat interface AI Integration: Custom prompt engineering to enable structured function calling and context-aware responses External Services: Google Calendar and Google Sheets APIs with custom wrappers for simplified interaction
Challenges we ran into
Creating a reliable function calling system that could translate natural language into precise API operations Managing timezone conversions correctly when working with calendar events Building a robust error handling system that communicates failures clearly to users Designing an architecture that could be extended to support additional integrations beyond the initial set
Accomplishments that we're proud of
Developed a clean, intuitive interface that makes complex operations feel simple Created a flexible backend architecture that can easily incorporate new external services Built a system that truly bridges the gap between conversation and action Implemented proper timezone handling to ensure calendar events appear correctly
What we learned
The importance of designing prompt engineering specifically for function calling How to build a backend that can reliably translate between natural language requests and API operations Techniques for robust error handling in a complex, multi-service system The challenges of building a cohesive UX across natural language and structured data
What's next for toliq
Adding more integrations with productivity tools like Notion, Trello, Slack, and email services Implementing more complex workflows that can span multiple services Developing a memory system to better understand user preferences and common tasks Creating a mobile application for on-the-go productivity Adding authentication and multi-user support for team environments
Log in or sign up for Devpost to join the conversation.