⭐ If you find this project useful, please give it a Star on GitHub! It helps the project grow.
A lightweight Python AI engine built from scratch with an interactive Flask Web Hub UI, live self-learning, multi-query processing, dynamic math solving, and automated JSON memory management.
-
🌐 Interactive Web Interface (
app.py): Modern Flask UI featuring live model switching between v7.0 Neural and v5.6 Legacy, real-time JSON memory selection, dynamic engine notices, and automated cleanup of empty memory files. -
🧠 Live Interactive Training: If the AI doesn't understand a query, it interactively asks for the correct response and updates its persistent memory instantly.
-
⚡ Auto-Fallback Engine: Seamlessly detects local dependencies such as TensorFlow and safely falls back between v7.0 and v5.6 without freezing the session.
-
🧮 Dynamic Math Solver: Instantly evaluates arithmetic expressions embedded directly within text queries.
-
🔀 Multi-Sentence Splitting: Breaks down complex user inputs joined by commas or connectors and processes the queries sequentially.
-
🤝 Smart Memory Merging: Safely merges training data from multiple JSON files into a master memory without data loss.
| File Name | Purpose | Key Concept |
|---|---|---|
app.py |
Web Hub Server | Flask server powering the Web Hub UI, model switching, JSON memory selection, and automatic cleanup. |
v5_6_AI_engine.py |
v5.6 Legacy Engine | Fast, lightweight rule-based engine with live training and arithmetic solving. |
main_engine.py |
v7.0 Neural Engine | TensorFlow-powered engine for advanced intent classification. |
merge_memory.py |
Memory Merger | Safely merges multi-user JSON training files into a master memory file. |
requirements.txt |
Python Dependencies | Lists the external Python packages required by Infinity AI Hub. |
ai_memory_*.json |
Persistent AI Memory | JSON databases storing trained patterns, intents, and custom user responses |
LICENSE |
Project License | MIT License information |
knowledge.db |
Offline Knowledge Base | SQLite database storing pre-built facts, fallback responses, and general knowledge queries. |
Open your terminal in the project directory and run:
python app.pyThen open your browser and navigate to:
Use the header dropdown to switch between:
- v7.0 — Neural Engine
- v5.6 — Legacy Engine
You can change the active engine directly from the Web Hub.
Switch between available .json memory files instantly without restarting the server.
Click ➕ New JSON to create a fresh user memory/session.
The Web Hub automatically scans memory files and cleans up empty memory files when required.
Chat naturally with the AI.
If an input is unrecognized, the AI will ask:
"Mujhe iska matlab nahi pata... toh kya jawab doon?"
Simply type the response you want the AI to learn.
The response is then saved permanently into the currently active .json memory file.
This allows the AI to continuously expand its knowledge through interaction.
The AI can detect arithmetic expressions directly inside normal text queries and calculate them automatically.
For example:
What is 25 + 75?
The engine can process the mathematical expression and return the result without requiring a separate calculator mode.
The engine can split complex user inputs into multiple smaller queries.
This allows the AI to process several requests from a single user message instead of treating the entire message as one query.
Example:
Hello, what is AI, how are you?
The engine can process the individual parts sequentially.
Infinity AI Hub supports two engine versions:
- v7.0 Neural Engine
- v5.6 Legacy Engine
v7.0 Neural Engine
│
▼
TensorFlow Available?
┌────┴────┐
YES NO
│ │
▼ ▼
v7.0 v5.6
Neural Legacy
If required local dependencies such as TensorFlow are unavailable, the system can safely fall back to the v5.6 Legacy Engine.
This helps keep the chat session running instead of failing because of a missing dependency.
To combine training data from multiple JSON memory files, place the exported memory files in the project directory and run:
python merge_memory.pyThe memory merger combines the training data into the master memory while protecting existing priority data.
┌─────────────────┐
│ Run app.py │
└────────┬────────┘
│
▼
┌──────────────────────────┐
│ Open 127.0.0.1:5000 │
│ in Browser │
└────────────┬─────────────┘
│
▼
┌──────────────────────────┐
│ Select Engine │
│ v7.0 Neural / v5.6 Legacy│
└────────────┬─────────────┘
│
▼
┌──────────────────────────┐
│ Select Active Memory File │
└────────────┬─────────────┘
│
▼
┌───────────────┐
│ Chat with AI │
└───────┬───────┘
│
┌──────────────┼──────────────┐
│ │ │
▼ ▼ ▼
Math Query Known Query Unknown Query
│ │ │
▼ ▼ ▼
Instant Answer AI Response Ask for Answer
│
▼
User Provides Response
│
▼
Response Saved to Memory
│
▼
AI Learns
│
▼
Merge JSONs When Needed
│
▼
python merge_memory.py
Install the required Python packages from requirements.txt:
pip install -r requirements.txtpython app.pyhttp://127.0.0.1:5000
v7.0 Neural
or
v5.6 Legacy
Choose an existing .json memory file or create a new one.
Ask questions, solve math problems, and train the AI with new responses.
When you have training data from multiple users:
python merge_memory.py ┌─────────────────────┐
│ Infinity AI Hub │
└──────────┬──────────┘
│
▼
┌─────────────────────┐
│ app.py │
│ Flask Web Hub │
└──────────┬──────────┘
│
┌──────────────┴──────────────┐
│ │
▼ ▼
┌─────────────────────┐ ┌─────────────────────┐
│ v7.0 Neural │ │ v5.6 Legacy │
│ main_engine.py │ │ v5_6_AI_engine.py │
└──────────┬──────────┘ └──────────┬──────────┘
│ │
└──────────────┬──────────────┘
│
▼
┌─────────────────────┐
│ ai_memory_*.json │
│ Persistent Memory │
└──────────┬──────────┘
▲
│
┌──────────┴──────────┐
│ merge_memory.py │
│ Memory Merger │
└─────────────────────┘
Infinity AI Hub is designed as a lightweight, self-learning conversational AI system that can:
- 💬 Chat with users
- 🧠 Learn from new conversations
- 💾 Store knowledge in JSON memory
- 📚 Retrieve offline factual data via SQLite (knowledge.db)
- 🔄 Switch between AI engine versions
- ⚡ Automatically fall back when dependencies are unavailable
- 🧮 Solve mathematical expressions
- 🔀 Process multiple queries
- 🤝 Merge knowledge from multiple users
- 📄 Extract and process text from PDF files
- 🖼️ Read text from images using OCR
- 🌐 Provide an easy-to-use Flask Web Hub
python app.pypython merge_memory.pyInfinity AI Hub is licensed under the MIT License - see the LICENSE file for details.
Copyright (c) 2026 infinityminds-dev
Infinity AI Hub is a lightweight AI project focused on experimentation, self-learning, persistent memory, and conversational AI development.
Built from scratch with the goal of creating a system that can continuously learn, remember, and improve through interaction.
If you like Infinity AI Hub and want to support its development:
- Give this repository a Star ⭐️
- Share it with other developers 🚀
- Report bugs or suggest features via Issues 💡