Launching today
DescribX Alpha
Master English expression by describing photos you take
2 followers
Master English expression by describing photos you take
2 followers
DescribX helps non-native English speakers smash the fluency plateau using multimodal AI. Snap or upload any real-world photo, speak or write your description, and receive a granular evaluation mapped to the CEFR framework alongside explicit pronunciation subscores and advanced vocabulary alternatives. Also features AI-driven Shadow Learning with hyper-realistic human voice generation and offline PDF/ZIP bundle exports. 100% free with no storage of your photos, voice recordings or descriptions.
Hey Product Hunt! 👋
I’m the solo developer behind DescribX.
As a non-native English speaker, I noticed a common wall language learners hit: we master textbook grammar, but when we look at the real world around us, we still struggle to describe everyday items or scenes fluently and expressively.
As a developer, I am completely fine using English in the workplace because we use a shared technical vocabulary—reviewing code, troubleshooting issues, etc. However, outside of my job, I often felt helpless describing the physical world around me, falling back on broken sentences, simple words, and basic structures.
When I looked for help, I found there is a lack of free, accessible tools that help English learners improve their English expression skills. I wanted a way to practice anytime, anywhere, without the friction or social anxiety of booking a tutor or talking to strangers (being someone with C-PTSD my entire life makes this even more difficult if not impossible).
So, I built DescribX to practice describing what I eat at the dinner table, what I encounter walking down a backstreet, or a masterpiece of art on a wall.
🚀 What DescribX Gives You
What DescribX is NOT: DescribX is not another digital English teacher offering rigid lessons and templates, nor is it a basic editing tool designed to just find and correct your spelling or grammar errors.
What it actually is: It is a professional, multi-dimensional assessment and shadow-learning engine built to bridge the gap between textbook rule-following and expressive, real-world fluency.
1. Image Description & Instant Context-Aware CEFR-Based Feedback
• Capture: Snap a photo anywhere, anytime, or upload an existing image.
• Describe: Speak or write your description of that exact image.
• Analyze: Get a professional assessment across 4 core dimensions: Vocabulary & Grammar, Accuracy, Cohesion & Coherence, and Task Fulfillment. If you speak, you also get granular pronunciation subscores.
• Level Up: The engine returns advanced phrasing alternatives tailored exactly one level higher than your current baseline along the six-scale CEFR framework.
• Export: Download everything into a clean PDF for offline review (includes the image, your description/transcript, and the full multi-dimensional assessment).
2. AI-Driven Shadow Learning
• Expert Generation: Snap or Upload any photo and ask the application to automatically generate an expert-level descriptive text.
• Realistic Audio: The system generates a realistic, human-like audio file of the description so you can listen, mimic, and shadow-practice.
• Export: Download everything as a single ZIP file containing the image, the generated text, and the high-quality audio file.
🎯 Who We Built This For
• High-Stakes Test Takers: Specifically designed to help candidates ace the "Speak/Write about the photo" tasks in the Duolingo English Test (DET), or the "Describe Image" or similar sections in PTE, IELTS, TOEFL, and Cambridge exams. DescribX cures performance blindness and reduces test anxiety by providing explicit overall scores and subscores before test day.
• Business Professionals: Built for professionals who struggle to present data, describe infographics, or present slides without frequent hesitations or "clutter" words. By practicing with charts and business imagery, you can master prosody and structural logic to turn every presentation into a demonstration of clarity and competence.
• Global Content Creators & Influencers: For creators who want to expand their global audience but worry about a stiff or unnatural verbal delivery. By utilizing our Shadow Learning and Pronunciation Assessment, you can master the natural "music" of the language, gaining the confidence to speak with charismatic, native-like authority.
• Global English Learners: Anyone who wants to break past the intermediate fluency plateau and articulate the world around them with professional precision.
🛠️ The Tech Stack Under the Hood
For the curious builders here, DescribX is a minimalist full-stack JavaScript application:
• Frontend: React, Javascript/HTML/Tailwind CSS.
• Backend: Node.js deployed entirely on Azure.
• Core Engine: Multimodal AI pipelines process the visual context of your photo to evaluate the exact accuracy of your text description or real-time audio stream.
💡 How to Try It Right Now
You can test the core features directly on the homepage with zero account creation (up to 20 free credits per day):
1. Try the standalone voice input to check your pronunciation accuracy.
2. Upload an image, type your description, and receive a full assessment.
3. Upload an image and generate an expert-level description.
Note: To unlock unlimited daily credits, evaluate your pronunciation side-by-side with your text assessment, or generate realistic human audio for shadow learning, simply create a free account to log in. I've kept it completely free during our alpha phase to ensure you can test it to its limits!
I am acting as a solo developer, architect, and ops team for this project. I would highly appreciate any feedback on latency, UX layout, or the scoring engine itself!
Check it out here:
https://www.describx.io