AIJudge

Company: AIJudge

Self-reported purpose: To transform documents into realistic speaking practice with an AI evaluator.

Key claim: Users can upload PDFs, PPTX, or DOCX files and receive voice-based practice sessions that evaluate understanding, logic, delivery, and persuasion.

What changed: The project description indicates a shift from traditional study tools to interactive, AI-driven oral practice — including voice recognition, semantic evaluation, and personalized feedback.

Most important open question: Is there evidence of user adoption or traction beyond the author's own development experience?

Analysis basis: Self-reported only. No archived data, revenue, customers, or third-party validation. All claims are from the author’s description.

The description states that AIJudge is a tool that:

• Accepts PDF, PPTX, or DOCX documents.
• Converts them into speaking practice sessions.
• Uses voice input for answers and evaluates performance across multiple dimensions (understanding, logic, delivery, persuasion).
• Provides feedback on response time, pauses, rhythm, filler words, and speech-to-text evidence.
• Offers five distinct evaluator personalities (e.g., Lumi Coach, Pressure Interviewer, Cody Teacher, Judge).
• Supports both Korean and English.
• Integrates GPT-5.6 for semantic evaluation and Web Speech API for voice recognition.

Inference: The product appears to be a document-aware speaking practice platform using AI for both content understanding and performance feedback. It is not a general-purpose AI assistant or chatbot but a specialized tool for oral communication training.

The author states:

• Traditional study tools focus on reading, memorization, or static quizzes.
• AIJudge aims to provide a practice partner that listens, challenges reasoning, and shows how to improve.
• The tool supports both text and voice input, with structured feedback.
• Feedback includes category scores, explanations, model answers, and follow-up questions.

Claim: AIJudge positions itself as an advanced oral practice tool for learners who want to improve speaking skills in a realistic, AI-evaluated environment.

Inference: The positioning evolved from a general-purpose educational tool to one focused on speaking fluency and communication confidence, using document context to tailor practice.

The description states:

• Users upload documents (PDF, PPTX, DOCX) for practice.
• Practice is tailored to the content of the document.
• Feedback is aimed at improving speaking ability in contexts like interviews, presentations, or oral exams.
• The tool supports Korean and English.

Inference: The target customer likely includes students preparing for oral exams, professionals practicing interview skills, or individuals seeking confidence-building in speaking.

Not evidenced: No explicit segmentation of users by profession, age group, or learning level.

The description does not state:

• Whether the tool is free to use or paid.
• How revenue would be generated (e.g., subscriptions, one-time purchase, enterprise licensing).
• If pricing tiers exist or if monetization is planned.

Not evidenced: No business model or pricing information provided.

The description states:

• Built with React, TypeScript, Vite.
• Uses OpenAI API (GPT-5.6) for semantic evaluation.
• Voice recognition via Web Speech API.
• Serverless route on Vercel to manage API keys securely.
• Rule-based measurements for timing, pauses, rhythm, and filler words.
• Supports multilingual input (Korean, English).
• Document classification by type to generate relevant questions.

Inference: The tool is a web-based application with AI integration for content understanding and voice analysis. It uses modern frontend and backend stacks, and integrates with OpenAI APIs for core functionality.

The description states:

• The project was submitted to the OpenAI 2026 hackathon.
• It was built by one person (Hye-Min Jeong) with help from ChatGPT and Codex.
• No mention of users, customers, or usage metrics.
• No evidence of revenue, retention, or product-market fit.

Not evidenced: No traction data, user base, or adoption metrics. The project is described as a hackathon submission, not a commercial product in use.

The description does not mention:

• Direct competitors.
• Similar tools in the market for speaking practice or AI-assisted learning.
• How AIJudge differentiates from existing platforms.

Not evidenced: No competitive landscape or differentiation analysis provided.

Key risks inferred from the description:

• The tool is described as a hackathon project with no commercial traction or user feedback.
• It relies on GPT-5.6 and Web Speech API, which may not scale or be reliable in production.
• Voice recognition accuracy may vary across browsers and languages.
• The author has no software development experience — raises questions about long-term maintainability or scalability.
• No evidence of monetization strategy or business model.

Red flag: Lack of user data, revenue, or product-market fit makes it difficult to assess commercial viability.

1. What is the intended user base and how do you plan to reach them?
2. How will you monetize this tool? Is there a pricing model in mind?
3. Have you tested the tool with real users beyond the development phase?
4. How do you plan to scale voice recognition and semantic evaluation for broader use?
5. What are your plans for multilingual support beyond Korean and English?
6. How will you ensure consistent performance across different browsers and devices?
7. Are there any legal or privacy concerns around speech data collection and storage?