ai-projects-hub
Health Pass
- License — License: MIT
- Description — Repository has a description
- Active repo — Last push 0 days ago
- Community trust — 10 GitHub stars
Code Pass
- Code scan — Scanned 12 files during light audit, no dangerous patterns found
Permissions Pass
- Permissions — No dangerous permissions requested
This repository is a collection of 20 beginner-friendly AI projects and tutorials, primarily focused on building agents and automation tools using Python and the Claude API. It provides runnable code for tasks like PDF analysis and stock market tracking.
Security Assessment:
Overall Risk: Medium. The light code audit found no dangerous patterns, malicious code, or hardcoded secrets. However, by design, these projects make external network requests (e.g., to the Anthropic API, financial data APIs, and Twilio). Users must supply their own Claude API keys, so safe secret management (such as using environment variables) falls entirely on the developer. The projects also handle user inputs like PDFs and SMS messages, which are common attack vectors if deployed publicly. Treat this as a starting point requiring proper security hardening before any production deployment.
Quality Assessment:
The project is actively maintained, with the most recent push occurring today. It is completely open-source under the permissive and standard MIT license. Community trust is currently low but growing, as it has just started gaining traction with 10 GitHub stars. The repository is well-organized, contains clear documentation, and provides straightforward installation instructions.
Verdict:
Safe to use for local development and learning, but use with caution and implement strict security practices if deploying the apps publicly.
20 real-world AI projects with runnable code — RAG, agents, automation, and India-first tools using Python and LLMs.
I built this because most AI tutorials either don't run or teach you nothing real.
Every project here is something I actually wanted to use — and most are India-first
because tools like these almost always ignore ₹ and NSE.
⚡ Get running in 3 commands
git clone https://github.com/mahak867/ai-projects-hub.git
cd ai-projects-hub/01-pdf-chat-claude
pip install -r requirements.txt && streamlit run app.py
You need Python 3.10+ and a free Claude API key → console.anthropic.com
📦 Projects
🟢 Beginner — start here
01 · PDF Chat with ClaudeUpload any PDF, ask questions in plain English. Claude answers with exact page references — not hallucinated ones.
|
02 · Indian Stock AgentAsk plain-English questions about NSE/BSE stocks. Claude fetches real data via tool use and gives structured analysis.
|
03 · WhatsApp AI BotClaude on WhatsApp via Twilio. Full conversation memory per phone number. Deploy to any server.
|
04 · YouTube SummarizerPaste a YouTube URL → structured notes with key points, quotes, and action items. Works with auto-generated subtitles.
|
05 · Resume AnalyzerUpload resume + paste job description → match score out of 100, missing skills, section-by-section rewrite suggestions, cover letter points, and interview prep topics.
|
🟡 Intermediate
06 · NSE Earnings MonitorWatches NSE/BSE earnings announcements, runs Claude analysis on each result, fires Telegram alerts with EPS beat/miss summary.
|
07 · Claude MCP Financial AnalystMCP server for Claude Desktop. Gives Claude live NSE/BSE data, stock screening, and portfolio math — usable in every conversation.
|
08 · Multi-Agent Research PipelineThree agents in sequence: Researcher gathers data → Critic finds gaps → Writer synthesises. Better output than any single prompt.
|
09 · RAG over Annual ReportsIngest NSE annual report PDFs, ask questions across multiple years, get answers with page-level citations.
|
10 · Computer Use DemoWorking examples of Claude's computer use API — screenshot analysis, bash execution, and file editing in a safe sandbox.
|
11 · AI Code ReviewerGitHub Action that reviews every PR with Claude. Posts structured feedback with severity levels. 2-minute setup on any repo.
|
12 · Voice Journal AIRecord your day as voice or text → Claude extracts mood score, wins, todos, recurring themes → saves structured JSON → 30-day trend analysis.
|
🔴 Advanced
13 · Claude Trading SignalsComputes RSI, MACD, Bollinger Bands with pandas → Claude interprets the pattern → BUY/SELL/HOLD with entry, stop loss, target, and R:R ratio.
|
14 · Build Your Own PerplexityWeb search (Brave API, free tier) + Claude streaming = cited AI answers that update in real time. Full conversation history.
|
15 · Agentic Data AnalystUpload any CSV → Claude writes pandas analysis code tailored to your data → executes it in a sandbox → explains results in plain English.
|
16 · Real-Time News AnalystPulls ET, MoneyControl, Reuters, Bloomberg RSS → Claude finds cross-source themes → daily email digest with investment thesis and risk flags.
|
17 · Open Source ScreenerSelf-hosted Screener.in replica. P&L, Balance Sheet, Cash Flow, Shareholding, Analyst Ratings. Claude AI analysis panel for Pro users.
|
18 · Document IntelligenceExtract structured data from PDFs using a JSON schema. Compare multiple documents — e.g. 3 years of annual reports — side by side.
|
19 · AI Interview CoachMock interviews for any role and difficulty level. Every answer gets a score out of 10, a grade, specific feedback, and a sample strong answer.
|
20 · Context Engineering Cookbook10 prompting patterns that work in production — structured output, self-critique loops, constitutional constraints, retrieval formatting — each with runnable code.
|
🏗️ Structure
Every folder has exactly three things — no more, no less:
project-name/
├── main_script.py ← working code, no stubs
├── requirements.txt ← exact deps
└── README.md ← setup, how it works, what to learn
🇮🇳 Why India-first?
Projects 02, 06, 07, 09, 13, 16, 17 are built specifically around NSE/BSE data, ₹ denominations, and Indian market structure. The majority of AI finance repos treat the US market as default and everything else as an afterthought. These don't.
🤝 Contributing
Found a bug or want to add a project? See CONTRIBUTING.md.
Issues tagged good first issue are a good place to start.
Reviews (0)
Sign in to leave a review.
Leave a reviewNo results found