Mavin

• 

  A conversation with an audio input with output in hinglish along with an hindi audio.

Inspiration

The inspiration for Mavin came from the need to streamline and automate the labor-intensive process of creating research reports. I envisioned a tool that could handle multimodal inputs, translate content seamlessly, and produce high-quality outputs, all while integrating effortlessly with popular platforms like WhatsApp.

What it does

Mavin automates the entire process of generating research reports. It breaks down complex objectives into manageable sub-tasks, transcribes audio, analyzes images, translates content into multiple languages, and converts Markdown to beautifully styled PDFs. It leverages AI orchestration and sub-agents to ensure accuracy and efficiency, making it an indispensable tool for researchers and professionals.

How we built it

I built Mavin using Vertex AI's Gemini models for orchestration and sub-agent tasks, Google Cloud's speech and translation services for audio and language processing, and advanced PDF handling tools for document generation. The system integrates various APIs and services to manage workflows, store files securely, and deliver high-quality reports.

Challenges we ran into

One of the main challenges was ensuring the seamless integration of various AI services and maintaining accuracy across different modalities (text, audio). We also faced difficulties in optimizing the workflow to handle large volumes of data efficiently and ensuring the translation accuracy for different languages.

Accomplishments that we're proud of

We are proud of successfully creating a robust system that automates the complex process of research report generation. Mavin's ability to handle multimodal inputs, perform accurate translations, and produce high-quality PDFs is a significant achievement. Additionally, integrating these features with WhatsApp for easy sharing and collaboration is a milestone we are excited about.

What we learned

Throughout this project, we learned the importance of modular design in building scalable and efficient AI solutions. We also gained deeper insights into the capabilities and limitations of various AI models and services, and how to leverage them effectively to solve real-world problems

What's next for Mavin

Next, we plan to enhance Mavin's capabilities by incorporating more advanced AI models and expanding its language support. We aim to improve its integration with other platforms and tools to make it even more versatile. Additionally, we plan to add features like real-time collaboration and more customization options for report generation to further meet the needs of our users.

Built With

• gemini
• stt
• tts
• vertexai