🎙 Meeting Task Extractor

This Meeting Task Extractor converts meeting recordings or manual transcripts into actionable task lists.
It performs two key operations:

✅ Transcribes Audio (using Whisper large)
✅ Extracts Actionable Tasks (using Mistral-7B API)

The result is a neatly formatted JSON summary of responsibilities, ready for follow-up and delegation.

---

🌟 Features

• 🔉 Audio-to-Text Transcription (using Whisper large).
• 📝 Manual text fallback if you don't have audio.
• 🧑‍🏫 Actionable task extraction powered by OpenRouter's Mistral-7B API.
• 🔹 User-friendly Gradio UI.
• ✅ Ideal for meeting minutes, follow-up emails, and team task allocation.

---

🛠 Tech Stack

• OpenAI Whisper: Audio transcription
• OpenRouter (Mistral-7B API): Large Language Model for task extraction
• Gradio: User Interface for file upload or manual paste
• pydub: Audio format conversion (MP3 to WAV)

---

🔹 Installation

pip install -q gradio openai-whisper pydub ffmpeg-python

---

🔹 API Token

✅ Signup at OpenRouter to generate your API key. Save it to OPENROUTER_API_KEY.

OPENROUTER_API_KEY = "your_api_key_here"

---

🔹 Run

python app.py

➥ The Gradio UI will launch in your browser. ➥ You can upload an audio file or paste meeting minutes manually.

---

🔹 File Format Support

✅ Audio: .wav or .mp3. ✅ Text: manually pasted into the textbox.

---

🔹 How It Works

1️⃣ If you upload audio, Whisper converts it into text. 2️⃣ If you paste text manually, it directly uses it. 3️⃣ The transcript is then fed into OpenRouter's Mistral-7B API, which parses it for actionable tasks. 4️⃣ The output is displayed in JSON format, with fields:

[
  {
    "assigned_to": "Person's name",
    "department": "Team or department",
    "task_description": "Brief description of task",
    "deadline": "Due by (if available)"
  }
]

---

🔹 Run in Gradio UI

1️⃣ Open the link provided by Gradio in your terminal. 2️⃣ Option 1: Upload an audio file. 3️⃣ Option 2: Or paste your meeting's transcript manually. 4️⃣ Hit Process and view the results immediately.

---

🔹 Notes

➥ OpenAI Whisper large performs well but may be slow on weak machines. ➥ The OpenRouter API key must be kept private. ➥ The model (Mistral-7B) is lightweight, fast, and well-suited for this use case.

---

🔹 Possible Applications

✅ Automate meeting minutes ✅ Provide clear follow-up instructions ✅ Keep stakeholders informed about responsibilities ✅ Boost productivity and collaboration within your team

---

📝 License

This project is licensed under MIT license — feel free to reuse and modify it.