Generate structured, professional, and downloadable Minutes of Meeting (MoM) from handwritten notes, PDFs, Word Docs, Text files, and Images — all powered by Google Gemini 1.5 Flash and LangChain.
✅ Ideal for Construction, Civil, MEP, Electrical, and Project Management Teams
✅ Converts raw unstructured notes to standardized MoM table
✅ Exports to Word and (optionally) PDF with Summary + To-Do List
- 🧠 Gemini Vision + LangChain for contextual understanding of notes
- 📝 Accepts multiple input types: JPG, PNG, PDF, DOCX, TXT
- 🏗️ Structured output with:
- Work Area
- Sub-Activity / Component
- Floor / Section
- Description
- Assigned To
- Deadline
- Status / Completion %
- 📤 Download as .docx with Summary and To-Dos
- 🧼 Clean Streamlit interface with user instructions
Work Area | Sub-Activity | Floor | Remarks | Assigned To | Deadline | Status |
---|---|---|---|---|---|---|
Fire | Dismantling Shaft | 6th Floor | Remove shaft partition | Not Mentioned | 22/05/2025 | Completed |
Plumbing | Handover | 6th Floor | Handover done | Vendor A | 21/05/2025 | Completed |
✔️ Summary
✔️ Key Action Items
✔️ Auto-generated To-Do List
Streamlit
– UILangChain
– Prompt orchestrationGoogle Generative AI
– Text & Vision model (gemini-1.5-flash
)python-docx
– Word file generationdotenv
– Secure API key management
├── app2.py # Streamlit App
├── generate_mom.py # LangChain + Gemini Logic
├── formatting.py # Word export logic
├── .env # Stores your Gemini API Key
├── requirements.txt # Python dependencies
├── README.md # Project documentation
- Clone this repo
git clone https://github.com/yourusername/ai-mom-generator.git cd ai-mom-generator
- Install Dependencies
pip install -r requirements.txt
- Create
.env
file and add your Gemini API key:gemini_api_key=your_google_api_key_here
- Run the Streamlit App:
streamlit run app2.py
📌 Example Use Case 🛠️ A construction project team conducts an on-site review. They jot down handwritten notes on tasks like plumbing, shaft wall removal, and electrical ducting. 📸 They upload a photo of the notes. 🧠 The app reads it via Gemini + LangChain and generates a clean, downloadable Word document with tasks, status, and deadlines.
📃 License This project is open-source under the MIT License.
🙌 Credits Built by Mukul Chauhan, AI Strategist & Project Consultant.
💡 Future Enhancements
✅ Upload multiple files together (multi-modal ingestion) ✅ Multi-language support for Hindi / regional scripts ✅ Dashboard with filters by Work Area, Assigned To ✅ PDF + Word export in 1 click
Let me know if you want this README.md
as a downloadable file or auto-uploaded to your GitHub.