Smart Tube
Analyse your weakness in your YouTube journey
Created on 29th November 2025
•
Smart Tube
Analyse your weakness in your YouTube journey
The problem Smart Tube solves
Content Creator Helper bridges the gap between raw data and actionable strategy for aspiring YouTubers.
Most beginner creators struggle because they treat content creation as a guessing game. They upload videos, see low views, and get discouraged because they don't understand why it failed. YouTube Studio provides graphs, but it doesn't tell a beginner what to do next.
How it makes tasks easier:
Automated Auditing: Instead of spending hours manually comparing their video metrics to competitors, users get an instant AI analysis highlighting exactly what is working and what isn't.
Data Literacy: It translates complex metrics (tags, upload frequency, view velocity) into simple human language (Strengths & Weaknesses).
Creative Block: The new AI Video Idea Generator solves "blank page syndrome" by analyzing the channel's history to suggest 3 high-potential, difficulty-rated video concepts.
Objective Feedback: It acts as an unbiased consultant, removing emotional attachment from the analysis to give a fair optimization score (0-100).
Challenges we ran into
- Ensuring Consistent JSON Output from the LLM:
One of the biggest hurdles was getting the AI to return data that our frontend could reliably render as charts and lists. Initially, the model would sometimes return Markdown text around the JSON, or hallucinate keys.
Solution: I utilized the responseSchema and responseMimeType: "application/json" features in the Gemini SDK. This forced the model to adhere strictly to our TypeScript interfaces (AnalysisReport and VideoIdea), preventing frontend crashes during parsing. - Managing API Quotas & "Demo" Experience:
The YouTube Data API has strict quotas, and requiring a user to get their own API key immediately creates a high barrier to entry.
Solution: I built a dual-layer service architecture. I implemented a robust "Demo Mode" with realistic mock data that triggers the actual Gemini analysis. This allows judges and users to test the full AI capabilities without needing a YouTube API key, while still keeping the production-ready code for real channel fetching in youtubeService.ts. - Visualizing "Thinking" Time:
Since the analysis involves multiple steps (fetching data, processing metrics, generating insights), the wait time could feel long.
Solution: I moved away from a simple spinner and implemented a Granular Loading State. The UI now communicates exactly what is happening (Fetching Data -> Analyzing with Gemini -> Finalizing Report), which keeps the user engaged and trusting of the process.
Tracks Applied (3)
Best Use of GitHub
GitHub
The Apify $1M Challenge
Apify
Best Use of CodeRabbit
CodeRabbit.ai
Technologies used