EchoWhisper

EchoWhisper: Your lifelong AI companion — voice-powered, emotionally aware, and always there to guide, uplift, and connect. Speak. Feel & Grow.

Comment

Inspiration

We wanted to develop something larger than an ordinary AI chat app. We took inspiration from real-world experiences where people feel unheard, isolated, or are in need of emotional connection — especially for people who don't feel comfortable connecting directly in those moments. That was a part of the concept of EchoWhisper – an AI that listens, responds, and gives a person a safe digital space to speak or vent. We didn't just want to build a chatbot. We wanted to create an experience that felt human, warm, and non-judgmental , a balance of voice and text experiences. We wanted to create something that could really echo a person's experience and whisper hope back.

What it does

EchoWhisper is a conversational AI voice agent designed to act as a calming, emotionally intelligent companion. Users can speak to Echo, and it responds with empathetic, reflective dialogue — helping them feel heard and understood. It features:

Conversational Voice AI powered by ElevenLabs

Emotionally expressive personality modes like “Chill Guru” or “Motivator”

Smart mood-aligned responses that adapt in real-time

A beautifully designed neon-themed front-end UI

Easy onboarding without requiring typing or form-filling

It’s not a therapist — it’s a friendly listener with personality.

How we built it

Frontend: Built using React with custom CSS and animation to give a modern, neon “cyber” feel.

API: Integrated ElevenLabs Voice AI SDK for natural voice input and output

Voice Agent Setup: Configured with customized system prompts, voice models, and temperature tuning for emotional intelligence.

Deployment: We connected the AI agent using API keys, with full audio conversation flow.

Challenges we ran into

Over the course of the hackathon, we had technical challenges and platform challenges. For example,

Deployments randomly failed, even though the code seemed fine.

Code would randomly be eliminated or overwritten — sometimes just when saving — or importing.

We hit the token limits very quickly (in a few cases after only 2-3 prompts) which really caused slowdowns during critical times of development.

Importing projects from GitHub or from other folders (on the local machine) did not always go smooth, sometimes generating unknown errors.

Another big challenge we faced was that 4 of the 6 of us were not familiar at all with building and deploying projects like this, and it took time to learn how the platform worked, understand where the bugs were coming from, and learn how to collaborate effectively. But we all supported each other, we were patient, and we learned together – and that’s how we made this project come to life.

These challenges did not defeat us - they made our team stronger and more resolute to finish what we had started.

Then, API Limitations & Access: Figuring out the limits of ElevenLabs’ Creator Access and when LLMs start incurring costs was confusing.

Voice Agent Integration: Mapping voice inputs to accurate, emotionally tuned outputs took fine-tuning.

Model Behavior Control: Designing the right system prompts and personality layers was key, but tricky.

Frontend Coordination: Coordinating frontend animation, sound syncing, and UI testing under time pressure was difficult.

Accomplishments that we're proud of

Created a fully functional voice-based AI companion within a tight timeline.

Designed a standout futuristic UI with consistent branding and appeal.

Successfully integrated ElevenLabs to form natural, reflective conversations.

Developed a custom personality system for voice tone customization — a major highlight for user experience.

What we learned

Working with LLMs in real-time voice scenarios is highly nuanced.

Prompt engineering and tone/personality shaping are just as important as model power.

APIs and tokens are powerful, but require careful setup and cost-awareness.

Team coordination and version management are critical in hackathons.

What's next for EchoWhisper

Enable emotional tone detection using voice sentiment analysis.

Add a "Time Capsule" feature where users can send themselves messages for the future.

Release a mobile version for easier accessibility.

Launch with pre-built personalities users can switch between on the fly (e.g., Calm Coach, Cheerful Buddy, etc.)

Add a support AI chat from BharatGPT to support the users while they use it

Explore mental wellness partnerships to make EchoWhisper a safe space for more people.

Built With

Share this project:

Updates

posted an update —

As a fresher, embarking on my college journey in just a few months, this hackathon marks my very first experience, and I've poured immense effort into making it a reality. Initially, I underestimated the potential impact of this project. However, witnessing the numerous cases in Kota – students overwhelmed by stress stemming from loneliness and academic pressures – profoundly changed my perspective. This isn't an isolated issue; globally, reports consistently highlight how stressed students often develop health problems like elevated blood pressure and experience a profound sense of isolation and emotional distress. This is particularly challenging for teenagers, who often internalize their struggles and hesitate to share their experiences, fearing judgment. This realization became the driving force behind our app. Introducing 'Eco Whisperer': our app is a vital digital companion addressing the global crisis of student mental and physical well-being. We provide a judgment-free, empathetic space for students facing immense academic pressure, loneliness, and stress-related health issues. Why this matters (backed by research):

  • Mental Distress: 1 in 7 adolescents globally experience mental disorders.[1, 2, 3] In the U.S., 36% of college students reported anxiety, 41% depression.
  • Loneliness: Affects 17-21% of young adults worldwide [4]; 36% of Kota students feel lonely.
  • Academic Pressure: Impacts nearly 44% of students in high-stakes environments like Kota.[5]
  • Physical Health: Chronic stress can contribute to issues like high blood pressure, seen in 25-35% of Indian youth.[6]
  • Help-Seeking Barriers: Only 3% of Kota students seek professional help due to stigma. How 'Eco Whisperer' helps: Our AI-powered app offers companionship (Chill Companion, Supportive Friend), emotional support (Motivator, Chill Guru), and academic guidance (Voice Agent Echo) to bridge this gap. Important Disclaimer: While 'Eco Whisperer' is a supportive tool, AI is not a substitute for professional human mental health care. For severe concerns or crises, please seek immediate help from qualified professionals or emergency services. Check out our demo!

Log in or sign up for Devpost to join the conversation.