Turn AI project ideas into transparent, comparable, and actionable opportunity analysis.
IdeaScoreAgent is a local AI Skill & Business Opportunity Scoring Dashboard for evaluating AI Skills, AI Agents, content ideas, lightweight business opportunities, and portfolio projects—without an API key or external AI service.
All screenshots below were captured locally with the built-in, public-safe AI 作品集项目评分器 sample. No private user or customer data is shown.
Capture guidance and the public-data checklist are documented in docs/SCREENSHOTS_GUIDE.md.
The app collects a structured idea brief and applies a deterministic rule-based scoring engine. It returns:
- a 0–100 Total Score and opportunity level;
- ten business and execution dimension scores;
- a separate risk level and penalty;
- Target User, Pain Point, MVP, Monetization, and Portfolio analysis;
- a Day 1–Day 7 validation plan;
- downloadable Markdown and CSV results.
The rules are visible in modules/scoring.py; the same input produces the same result.
This project connects business judgment with practical AI product development. It shows how a business / Marketing background can contribute to AI work through user segmentation, pain-point analysis, willingness-to-pay assumptions, MVP scope, risk boundaries, and clear reporting—then turn those ideas into a tested Streamlit product suitable for GitHub, a résumé, interviews, and content creation.
- 0–100 Opportunity Score with Strong / Good / Medium / Weak levels
- Ten transparent scoring dimensions and four risk levels
- Keyword, completeness, buyer-type, feasibility, and complexity rules
- Five built-in, portfolio-relevant sample ideas
- Apple-inspired Streamlit interface with clear metric cards
- Dimension table with score interpretations
- Business analysis and visible risk warning
- Seven-day validation action plan
- Formal 11-section Markdown opportunity report
- CSV score export with one row per dimension
- Public-safe reproducible sample outputs
- Automated tests for scoring, risk, boundaries, and report structure
| Dimension | Business question |
|---|---|
| Pain Intensity | Is the problem strong, concrete, and costly enough to matter? |
| Frequency | Does it happen daily, weekly, repeatedly, or at meaningful volume? |
| Willingness to Pay | Is there a clear buyer, budget, and payment trigger? |
| MVP Feasibility | Can one useful input-to-output loop be built quickly? |
| Data Availability | Can safe, legal, and representative test data be obtained? |
| Differentiation | Is the workflow more specific than a generic ChatGPT prompt? |
| Portfolio Value | Does it demonstrate analysis, automation, UI, testing, and reporting? |
| Monetization Potential | Is there a credible subscription, per-use, service, or licensing path? |
| User Clarity | Is the target user defined by identity, context, and frequency? |
| Execution Fit | Does the scope match a business + AI Agent learner's current skills? |
Total Score = Average(10 dimension scores) × 10 − Risk Penalty
Risk penalties: Low 0, Medium 5, High 10, Very High 20. Final scores are constrained to 0–100.
The five included examples are:
- 留学生课堂录音和 PPT 自动整理工具
- 小红书 / 抖音评论痛点分析工具
- 中小企业客户邮件自动分类和回复草稿工具
- AI 作品集项目评分器
- 微信卖课引流自动化工具
They demonstrate education, creator research, SME workflow automation, portfolio planning, and course-sales lead management while making privacy and platform risks visible.
- Python 3.11
- Streamlit
- Pandas
- Pytest
- Watchdog (optional Streamlit development performance helper)
- Rule-based Scoring Engine
No OpenAI or ChatGPT API is connected in the current version.
cd ~/AIProjects/IdeaScoreAgent
python3.11 -m venv .venv
source .venv/bin/activate
python -m pip install --upgrade pip setuptools wheel
pip install -r requirements.txtcd ~/AIProjects/IdeaScoreAgent
source .venv/bin/activate
pytest
streamlit run app.pyOpen http://localhost:8501. Stop the local server with Control + C.
- Markdown: formal 11-section Idea Opportunity Report
- CSV: overall result plus each dimension score and interpretation
- Local outputs: generated at runtime in
outputs/and excluded from Git by default - Public sample outputs: reviewed built-in example files in
portfolio/sample_outputs/
Regenerate public samples with:
python scripts/generate_sample_outputs.py- Not investment, legal, medical, or other professional advice
- No automatic trading, payment, diagnosis, approval, or high-risk decision-making
- No unauthorized scraping, private-data collection, or platform-rule bypass
- No API Key required; the current app runs locally and does not upload form data
- Real user outputs in
outputs/are private by default - Users should not enter passwords, private chats, payment details, health records, or business secrets
Risk Level is a keyword-based reminder—not a legal conclusion or security audit. See docs/SAFETY_BOUNDARY.md.
- Structured idea input
- Ten-dimension rule-based scoring
- Four-level risk penalty
- Five sample ideas
- Business explanation and seven-day action plan
- Markdown and CSV exports
- Automated test suite
- Public-safe sample outputs and showcase documentation
- Final README screenshots
- Multi-idea comparison
- User-defined weights
- PDF export
The current score helps prioritize validation. It does not prove market demand, product-market fit, income potential, or commercial success.
IdeaScoreAgent/
├── app.py
├── modules/ # scoring, reports, and built-in samples
├── tests/ # automated scoring and report tests
├── scripts/ # reproducible public sample generation
├── docs/ # overview, safety, and screenshot guidance
├── portfolio/
│ ├── screenshots/ # reviewed README screenshots
│ └── sample_outputs/ # synthetic, public-safe example reports
├── outputs/ # local user reports; ignored except .gitkeep
├── README.md
├── requirements.txt
├── LICENSE
├── CHANGELOG.md
└── PUBLIC_SHOWCASE_MANIFEST.md
- Multi-idea comparison and ranking
- User-defined dimension weights and templates
- PDF report export
- Business Model Canvas generator
- Competitor and interview-evidence modules
- GitHub portfolio page generator
- Optional OpenAI / ChatGPT API enhanced analysis after privacy and cost review
This project uses a custom Portfolio License, not MIT or GPL. Viewing, running, studying, and educational demonstration are permitted; commercial use, redistribution, or resale requires prior written permission. See LICENSE for the full terms and warranty disclaimer.
This project is part of my local-first AI Agent portfolio and can be managed through AgentHubControlCenter, the central command center for agent manifests, safe actions, useful signals, workflow simulations, connector readiness, approval gates, and public-safe reporting.
IdeaScoreAgent is registered as an idea validation module in AgentHubControlCenter.