Angin AI

Inspiration

Its all started with my friend who is just went through break up, she was needing a lot of emotional support to overcome the breakup. many of time, I couldn't be present and be the best listener for her. that's inspire me to build this app. an app that you can use every time you feel needs to talk to someone.

What it does

Angin is your go-to AI friend when you need someone to talk to. You can call, share what's on your mind, and Angin will listen closely, support you emotionally, and offer gentle, thoughtful suggestions.

How we built it

Member of this project is me, Kiro and Chatgpt. I vibe code the entire app using Chatgpt and Kiro without prior experience in Mobile app and audio handling application. I learn and figuring out as I go, I try not to implement too many features.

Challenges we ran into

The biggest challenge was integrating multiple parts (audio, LLM, TTS, UI, environment variables) while working under a tight deadline and without prior full-stack experience in audio pipelines.

Accomplishments that we're proud of

The vision that I had in my head was, how smooth and gentle the tone going to be and how emotionally intelligent the output of the LLM, and thanks to Kiro, Angin could have therapist agent embedded in it.

What we learned

Through this process, I learn almost everything that is implemented in Angin, connecting STT, LLM responses, and TTS into one seemless experience. This project strengthened my technical problem solving and my ability to deliver a working AI product across frontend and backend.

What's next for Angin AI

Improving the user experience. for now, we need to click a button to talk and sent the input to Angin. it will be better if Angin could have voice auto detections, just like Chatgpt pro, where it can stop listening if you are stop talking for 5 second. the other improvement will be the waiting time for the user to get the output or result from Angin. it took around 6second for the TTS providing the output.

Built With

• elevenlabs
• expo-av
• expo-speech
• json
• openai
• react-native
• typescript