💾 Persistent Github GPT-4o Agent (LangGraph + Streamlit)

A powerful, privacy-conscious AI assistant powered by GitHub-hosted GPT-4o.
Designed with developers in mind, this agent offers persistent session memory, secure token authentication, and extendability for real-world applications.

---

🔐 Authentication via GitHub Token

• Secure login via GitHub token (user-provided).
• Each token is SHA-256 hashed to create a private identifier (user_hash) used to:
  • Track user sessions
  • Prevent token storage
  • Isolate chat histories across users

💡 You can generate a GitHub token from:
https://github.com/settings/tokens

---

💾 Persistent Memory via SQLite

• All chat messages are stored in a local database (chat_messages.db).
• Each user can create and manage multiple chat sessions.
• All message types (human, AI, function, tool, etc.) are serialized and deserialized correctly.

---

🧠 LangGraph Architecture

• Built using LangGraph's StateGraph for controlled message flow.
• Node function (process) invokes the model and updates memory.
• Designed to be extensible for:
  • Tool calling
  • Function execution
  • Retrieval-augmented generation

---

🧠 Highlights

Feature	Details
🛡️ Token Hashing	Prevents token leaks and supports session-based user isolation
🗃️ Session Management	Named session support with persistent conversations
🧠 GPT-4o Integration	Uses GitHub's hosted inference endpoint (https://models.github.ai/inference)
🔄 Tool & Function Ready	Agent already supports future extensions for tool-calling and more
❌ Invalid Token Check	Validates GitHub token via trial request before login
🔄 Logout Flow	Secure reset and return to auth gate

---

📂 Project Structure

├── main.py # Streamlit frontend with authentication and UI
├── agent.py # LangGraph logic and GPT-4o invocation flow
├── github_wrapper.py # Custom wrapper for GitHub-hosted GPT-4o API
├── chat_messages.db # SQLite database for chat/session persistence
├── requirements.txt # Python dependencies
└── README.md # Project overview and usage

---

⚙️ How to Use

1. Clone the repo:

git clone https://github.com/your-username/GitHub-GPT4o-Agent.git
cd GitHub-GPT4o-Agent

2. Install dependencies:

pip install -r requirements.txt

3. Run it:

streamlit run main.py

4. Provide your GitHub GPT-4o token at login

---

🔒 Security First

• GitHub tokens are not stored.
• All session control is hash-based and local.
• SQLite stores only messages, not secrets.

---

🧰 Tech Stack

• LangGraph & LangChain
• Streamlit
• SQLite (for state persistence)
• GitHub GPT-4o Inference API

---

📸 Previews

🧑‍💻 Author

Jayesh Suryawanshi

• 🧠 Python Developer | 💡 AI Tools Builder | 🌍 Data & Engineering Enthusiast
• 📫 LinkedIn

---

📃 License

MIT License