dcrebbin / ai-interview-trainer-frontend

AI interview trainer designed to help you succeed at technical interviews, frontend developed in Qwik

Up It Quest

Up It Quest is an AI powered interview trainer for technical interviews

DEMO: https://www.youtube.com/watch?v=ef2ivitjiBU

Backend: https://github.com/dcrebbin/ai-interview-trainer-backend

Technology:

• Frontend: Qwik

• Backend: Go Fiber

• Database: MySQL with GORM

• Vector Database: Pinecone (Not currently leveraged really)

Default Hosting:

These are the providers I personally use to deploy everything, ofc can be interchanged to your preferences, but these work.

• Frontend: Vercel (using VITE)

• Backend: Google Cloud Platform

• Database: PlanetScale

• Vector Database: Pinecone (Not currently leveraged really)

Integrated AI Services:

Note: Vertex isn't fully supported and requires manual regeneration of the auth token on the backend (very sus): Mostly bc the other providers are better, sorry google (would be worth it if you have heaps of gcp credits or a partnership tho)

Text Generation:

• OpenAi (GPT3.5, GPT4 etc) (GPT4 Turbo is the best of both worlds)

• Vertex (Palm, Gemini Pro etc) (Maybe ultra will be insane but idk)

Text to Speech:

• OpenAi (TTS 1, TTS 1 HD) (Best of both worlds)
• Vertex
• ElevenLabs (Most expensive and the slowest but the highest quality)
• Unreal Speech (Cheapest and the quickest but slightly uncanny valley)

Speech to Text:

Both are pretty decent, whisper's probably a bit cheaper tho

• OpenAi (Whisper 1)
• Vertex

Setup

1. npm i

2. Create a local.env to match the example.env

3. Setup and run the API

4. Start-up the frontend!

ai-interview-trainer-frontend's Issues

can you make it work with ollama?

Checkout ollama.ai -- it has a very similar api to openai, but is free.

what goes in .env?

VITE_API_URL=
VITE_ENVIRONMENT="production"
AUTH_SECRET=
GOOGLE_CLIENT_ID=
GOOGLE_SECRET=
API_KEY=
API_URL=
```

this isn't very helpful to me.