📚 EduBotAI+ — Your AI-Powered Study Companion 🚀

Winner – 1st Place at the IITM Azure Learning Series Hackathon

EduBotAI+ is an AI-powered smart learning assistant that transforms how students engage with study material. It supports PDFs, images, handwritten notes, and slides—using Azure's services to extract content. Leveraging NLP techniques, it performs text extraction,grammar correction, summarization, translation, and flashcard generation. With added voice-based input/output, EduBotAI+ simplifies complex concepts into interactive, easy-to-understand formats—making learning faster and more personalized. l

---

✨ Features

📥 Upload & Choose Operation

Users can upload PDFs, images, or audio, and choose from:

• 📝 Text Converter – Extracts text from documents and images using OCR.
• 🧠 Summary Creator – Cleans and summarizes large texts for quick understanding.
• 🎙️ Voice Translator – Converts speech to text, translates, and generates a subtitle video or audio.

---

🎤 Voice Input → Text Conversion

• Uses: Azure Speech-to-Text
• Supports: Direct microphone input or uploaded .wav/.mp3 files
• Output: Recognized transcript saved as voice_transcript.txt

---

🔁 Flashcard Generator

• Converts summarized content or notes into Q&A flashcards
• Extracts key concepts using:
  • Keyword analysis
  • Sentence similarity
  • Bloom’s taxonomy (optional)
• Output format: Readable list and downloadable .txt or .json

---

🗣️ Text-to-Speech Translator with Video

• Input: User-typed or extracted text
• Language Options:
  • English 🇬🇧
  • Tamil 🇮🇳
  • Hindi 🇮🇳
  • Telugu 🇮🇳
• Output:
  • 🎧 Audio (.mp3)
  • 🎞️ MP4 video with:
    • Subtitles in selected language
    • Custom or default background image

---

🧠 Technologies Used

🔧 Backend Stack

• Python
• Flask
  • Lightweight Python web framework used to handle routing, templating, and RESTful API services
• Azure Services:
  • Azure Speech Service (STT + TTS) – Free Tier
  • Azure Computer Vision OCR – Free Tier
  • Azure Blob Storage – For storing processed files
  • Azure Cosmos DB – For storing metadata and logs
  • Azure Translator - For translating text to speech - Free Tier

🌐 Frontend Stack

• HTML/CSS/JavaScript
• Responsive UI for uploads, operations, and results display

---

✅ How It Works – Workflow Breakdown

Step 1: Voice Input

• Users speak or upload an audio file
• Azure STT converts speech to text
• Saved as transcript for next steps

Step 2: OCR (Text Extraction)

• PDFs, scanned images, and handwritten notes are processed
• Azure OCR extracts raw text from uploaded files

Step 3: Text Cleaning & Summarization

• Removes noise (special characters, formatting issues)
• Outputs clean paragraphs
• Summarization done using OpenAI or rule-based logic

Step 4: Flashcard Generation

• Smart Q&A generated from clean text
• Exportable for revision or quiz practice

Step 5: TTS and Subtitle Video

• User text is converted to speech (TTS)
• Subtitle video generated with selected language and image
• Stored in Azure Blob Storage for later access/download

---

📁 File Storage & Output Management

All outputs including:

• Cleaned text
• Summaries
• Flashcards (.txt/.json)
• Transcripts
• MP4 videos with subtitles

...are automatically uploaded and organized in Azure Blob Storage, with metadata tracked in Azure Cosmos DB for easy access and audit.

---

🧪 Example Use Case

A student uploads a PDF chapter and selects “Summary Creator” → EduBotAI+ extracts and cleans the text, summarizes it, and generates flashcards. They can also listen to the summary in Hindi or generate a video with Tamil subtitles for revision.

---

🚀 Future Enhancements (Student-Focused)

🔎 1. Semantic Search on Summarized/Corrected Content Enable intelligent search using vector embeddings (e.g., SentenceTransformers):

• 🔍 Search similar summaries based on meaning, not just keywords
• ❓ Ask questions across multiple documents
• 🧠 Smart retrieval: “Find notes where cybersecurity attacks were mentioned”

🧠 2. LLM-Powered Summary QA (Question-Answering) Integrate with LLMs like OpenAI GPT or Anthropic Claude for enhanced interaction:

• 📝 "Summarize this. Now give 3 bullet points and a call to action."
• ❓ Ask clarifying questions on the summary
• ✅ Improve summary coherence and factual accuracy

🧪 3. Multimodal Input Handling Extend input types beyond PDF/Image/PPTX:

• 🎧 Audio Support (MP3, WAV):
  • Transcribe using Whisper / Google STT
  • Clean and summarize the transcript
• 📹 YouTube or Video Link Support:
  • Frame-by-frame OCR for visual content
  • Summarize lecture visuals/texts

---

⚙️ Installation

• Clone the repository
• Install the virtual environment and dependencies
• Train the model and run the flask Application

---

👥 Collaborators

Name	GitHub Username
Shobhana S	@Shobhanashankar
Gautham R	@gautham-here
Sri Ranjana C	@sriranjanac