Note: The project showcases integrations of LLMs with real-world productivity tools. This is an in-progress GenAI tool.
Startup Copilot is a generative AI-powered assistant for early-stage startup founders.
It helps with:
- 🧠 Idea validation (Google Trends + LLM)
- ⚙️ Workflow generation
- 📊 Business case analysis
- 🎯 Go-To-Market strategy
- 📣 Pitch deck drafting
All results can be optionally synced to Notion!
- Frontend: Gradio UI (tab-based)
- Backend: Modal Functions (GPU, torch, transformers)
- Models Used: Zephyr 7B, Mistral 7B
- APIs: Google Trends via PyTrends, Notion API
- Deployment: FastAPI wrapped in Modal's ASGI App
Clone the repo and add your .env file with these variables:
MODAL_BASE=https://your-modal-app.modal.run
NOTION_TOKEN=your-notion-secret
NOTION_DATABASE_ID1=your-notion-dbThen install dependencies:
pip install -r requirements.txtRun locally:
python app.py(Add 1–2 images or a small GIF of your Gradio UI here if possible)
- Modal deployment
- LLM-based modules
- Notion integration
- PDF Pitch Deck formatting
- User auth and history
- Deployment to HF Space or Vercel (planned)
I’m exploring ways to make GenAI truly useful for early-stage founders. This project is part of my self-driven learning, applying APIs, cloud, and LLMs in practical ways.
- Prompt engineering for multiple agent roles
- LLM model handling with Hugging Face + torch
- Modal app building (FastAPI + GPU infra)
- API integration (Google Trends, Notion)
- Frontend UX with Gradio
MIT