explaingit

nils0000shiyong/kuaida-ai-assistant

22KotlinAudience · generalComplexity · 2/5Setup · moderate

TLDR

An Android app that listens to job interview questions, transcribes them, and generates a spoken-style answer in seconds using your own AI API key, with your resume uploaded as personal context for tailored replies.

Mindmap

mindmap
  root((Kuaida))
    Core flow
      Record question
      Auto-transcribe
      Generate answer
      Read aloud
    Answer controls
      Shorter version
      Add example
      More casual tone
    Personal context
      Upload resume PDF
      Project documents
      Local storage only
    Tech stack
      Kotlin and Compose
      MVVM architecture
      Hilt injection
Click or tap to explore — scroll the page freely

Code map

Detail Auto

An interactive map of this repo's files and how they connect — its source is parsed live in your browser. Click Visualize to build it.

filefunction / class

Things people build with this

USE CASE 1

Use during a live job interview to record a question and read a tailored spoken answer within 5 seconds

USE CASE 2

Practice mock interviews at home by recording sample questions and reviewing AI-generated answers in session history

USE CASE 3

Upload your resume PDF so every generated answer references your real experience instead of generic examples

USE CASE 4

Track your AI API spending per response while using the interview assistant

Tech stack

KotlinJetpack ComposeHiltMVVM

Getting it running

Difficulty · moderate Time to first run · 30min

Requires your own API key from a supported AI provider, the app is not on any app store and must be built from source or sideloaded.

No license information is mentioned in the explanation.

In plain English

Kuaida is an Android app designed to help people prepare for job interviews, aimed at Chinese-speaking job seekers. When an interviewer asks you a question, you record yourself speaking it or type it in, and within a few seconds the app produces a structured, conversational answer you can read aloud or adapt on the spot. The README describes the target response time as five seconds from hearing the question to seeing the answer. The core input method is audio recording. During a practice or live interview session, you press a button, speak the question, and the app transcribes it automatically before generating a reply. Answers are written in a natural spoken style rather than formal prose. If the initial answer does not feel right, buttons let you request a shorter version, ask the model to add a concrete example, or make the tone more casual. The app connects to AI language model providers using an API key that you supply yourself, with support for multiple providers from China and abroad. The interface shows token usage and the estimated cost for each response so you can track spending directly. Your API key is stored locally on the device only. A personal knowledge base feature lets you upload your resume or project documents as PDFs. The app uses those files as context when generating answers, so responses reference your actual experience rather than generic examples. A history screen keeps a log of past question-and-answer sessions for later review. The codebase is Kotlin with Jetpack Compose for the interface, an MVVM architecture split across multiple feature modules, and Hilt for dependency injection. The app targets Android 8.0 and above. The README states the app is intended for personal use only and is not planned for any app store release.

Copy-paste prompts

Prompt 1
I want to add a feature to the Kuaida Android app that lets users rate each AI answer with a thumbs up or down and saves the rating to the session history, help me implement this in Kotlin with Jetpack Compose
Prompt 2
Show me how Kuaida's MVVM architecture handles the flow from audio recording to transcription to LLM API call using Hilt dependency injection
Prompt 3
Help me extend Kuaida to support a new AI provider, what interface do I need to implement and where in the codebase does it plug in?
Prompt 4
I want to build a similar interview-prep Android app. Walk me through setting up Jetpack Compose with MVVM and an LLM API call that returns a response in under 5 seconds
Prompt 5
Show me how to implement a local PDF-as-context pattern in an Android app that loads a resume and passes it to every API call, similar to Kuaida
Open on GitHub → Explain another repo

← nils0000shiyong on gitmyhub — every repo by this author, as a profile.

Verify against the repo before relying on details.