wrapper4ai 0.1.2

A lightweight multi-provider wrapper for LLM chat with history and token management.

Wrapper4AI

Multi-provider, pluggable LLM wrapper with token counting, history management, and seamless extensibility.

🔧 Designed for developers building chat-based AI tools with OpenAI, Gemini, Bedrock, DeepSeek, Anthropic, and more.

---

🚀 Features

• 🔌 Unified interface for multiple LLM providers
• 💬 Chat history tracking with token trimming
• 🔢 Token counting with Tiktoken for OpenAI models
• 📚 Extensible handler base for adding new models
• ✅ Testable client API with clean abstraction

---

🛠️ Installation

  pip install git+https://github.com/DKethan/Wrap4AI.git

Or if you package it to PyPI:

  pip install wrapper4AI

---

🧩 Supported Providers

• ✅ OpenAI (gpt-4o, gpt-3.5, ... all OpenAI language models )
• ✅ Google Gemini
• ✅ Amazon Bedrock
• ✅ DeepSeek
• ✅ Meta LLaMA
• ✅ HuggingFace Interface
• ✅ Anthropic Claude
• ✅ Perplexity AI
• ✅ HuggingFace

---

🔧 Usage

1. Connect to a provider

from wrapper4AI.wrap import connect

client = connect(provider="openai", model="gpt-4o", api_key="sk-xxx")

2. Basic Chat

response = client.chat("Tell me a joke.")
print(response)

3. With History

history = [
    {"role": "user", "content": "What is Python?"},
    {"role": "assistant", "content": "A programming language."},
    {"role": "user", "content": "Who created it?"}
]
print(client.chat_with_history(history))

---

🧪 Testing

  python -m tests.test_openai_001 # make sure to add your api key in the test file