I built an open-source alternative to Cluely – Real-time AI interview assistant that’s completely transparent

Exploring Transparency in AI Interview Assistance: Introducing an Open-Source Alternative to Cluely

In recent weeks, the emergence of AI tools claiming to facilitate discreet and seamless interview and meeting interactions has garnered significant attention. One such innovation, Cluely, positions itself as an “undetectable AI” assistant capable of providing real-time answers during conversations. While this approach offers intriguing possibilities, it also raises important questions about transparency, privacy, and control — especially given its closed-source architecture.

Motivated by these considerations, a developer has created an open-source, self-hosted alternative designed to prioritize transparency and user autonomy. This project aims to deliver a reliable, customizable AI assistant that elegantly balances functionality with privacy.

Overview of the Open-Source AI Interview Assistant

This open-source solution incorporates several core features:

• Real-Time Audio Transcription: Utilizing the faster-whisper engine, the system transcribes conversations instantly, ensuring prompts and responses are synchronized with live speech.

• Question Detection & Answering: Leveraging the OpenAI API, it identifies relevant questions during interactions and provides accurate, context-aware answers on-the-fly.

• User-Friendly Web Interface: A streamlined, clean web-based UI allows users to monitor transcriptions and AI responses seamlessly during meetings or interviews.

• Multi-Platform Compatibility: Compatible across Windows, Mac, and Linux, offering broad accessibility without the need for specialized hardware or software environments.

Key Principles and Advantages

Unlike proprietary solutions, this project emphasizes a commitment to transparency and user control. Its key differentiators include:

• Complete Open Source Access: Users can review and modify all aspects of the code, fostering trust and customization.

• Self-Hosting Design: All processing occurs locally on the user’s machine, ensuring sensitive audio data remains private and avoiding reliance on external servers.

• Privacy-Centric Approach: The system refrains from making claims about being “undetectable.” Instead, it empowers users with full oversight of data flow and security.

• Cost-Free Usage: No subscription fees or proprietary lock-ins; enthusiasts and professionals alike can deploy and adapt the tool freely.

• Customization Capabilities: Users can tailor AI prompts, adjust the UI, and extend functionalities to suit specific needs or workflows.

Underlying Technology Stack

The robust architecture comprises:

• A Python backend managing real-time WebSocket communication.
• Faster-whisper for fast, efficient speech-to-text transcription.
• OpenAI’s API to perform question detection and generate contextual answers.
• A minimal