Summary

This project offers a free, open-source backend for AI-driven interview preparation and real-time communication. Targeting developers building Spring Boot AI applications and students, it provides automated mock interviews, resume analysis, and intelligent agent interactions via a clear, deployable architecture.

How It Works

Built on Spring Boot 3 and Java 17 with Spring AI, it uses a modular monolith architecture supporting HTTP, SSE, and WebSocket. It integrates AI conversations, intelligent agents, real-time ASR/TTS, and mock interviews. MySQL, MongoDB, and Redis handle data persistence and state management. A multi-agent workflow orchestrated by LiteFlow manages complex interview scenarios, ensuring session recoverability.

Quick Start & Requirements

• Install/Run: Docker Compose enables one-click deployment of the full stack (MySQL, MongoDB, Redis, application).
• Prerequisites: Java 17, Maven 3.6.3+, Docker.
• Links: Demo video on Bilibili. Documentation via WeChat (JavaAndGo8888 or hqy2004hw).

Highlighted Details

• Resume-Driven Interviews: Parses PDFs to generate personalized interview questions, scores, and suggestions via a multi-agent system.
• Distributed Single-flight: Ensures AI calls execute once across instances, preventing duplicate costs and enabling replay.
• Long Session State Management: Uses MongoDB snapshots and Redis lazy loading for recoverable runtime states, supporting interruption recovery.
• Real-time ASR: WebSocket with Xunfei AST provides end-to-end transcription featuring advanced incremental de-duplication.
• Unified AI Model Access: Spring AI abstracts various models (DeepSeek, Xinghuo) via an OpenAI-compatible protocol.
• Agent System: Supports runtime agent configuration with SSE streaming output.

Maintenance & Community

Developed by lishuangqiang and programmer dolphin, the project is free and open-source. Documentation is accessible via WeChat. No specific community channels like Discord/Slack are listed.

Licensing & Compatibility

• License Type: MIT License.
• Compatibility: Permissive for commercial use, modification, and integration into closed-source applications.

Limitations & Caveats

The project retains development-phase configurations; security hardening and key desensitization are planned for subsequent stages before final public release, suggesting potential need for further security review for production deployment.