# 🔌 pyllmlib
[](https://pepy.tech/projects/pyllmlib)
[](https://pypi.org/project/pyllmlib/)
[](LICENSE)
[](https://github.com/yazirofi/pyllmlib)
---
<div style="
width:200px;
height:100px;
overflow:hidden;
display:flex;
justify-content:center;
align-items:center;
margin:auto;
position:relative;
top:0;
bottom:0;
left:0;
right:0;
border-radius:10px;
box-shadow:0 4px 8px rgba(0,0,0,0.2);
">
<img src="https://raw.githubusercontent.com/yazirofi/cdn/main/yazirofi.jpg"
style="width:100%; height:100%; object-fit:cover;"
alt="Cropped Image">
</div>
---
**pyllmlib** is a lightweight, provider-agnostic Python package that gives you **one simple interface** to work with multiple Large Language Model (LLM) APIs.
Stop juggling different SDKs and client libraries — whether it’s **OpenAI GPT**, **Google Gemini**, **Mistral AI**, or **Groq**, you write your code once and switch providers in seconds.
🎯 **Ideal for** developers who want to:
* Experiment with different LLMs quickly
* Build multi-provider AI applications
* Avoid vendor lock-in with a consistent API
---
## ✨ Features
* 🔌 **Unified API** — A single `generate()` function for all providers
* 🌐 **Multi-Provider Support** — OpenAI, Gemini, Mistral, Groq (with more coming soon)
* 🧠 **Consistent Message Format** — Same request style across providers
* 🔐 **Flexible Config** — Use env vars, inline setup, or config files
* 📦 **Minimal Dependencies** — Only needs `requests`
* 🔄 **Quick Provider Switching** — Change models with one line
* 🛡️ **Automatic Token Handling** — Prevents overflows & context errors
* 📜 **Role-Based Conversations** — System, user, assistant messages
* 🔧 **Extensible** — Add your own providers with minimal code
* 🚀 **No Vendor Lock-In** — Swap providers without rewriting logic
---
## 📦 Installation
From PyPI (recommended):
```bash
pip install pyllmlib
```
From GitHub:
```bash
# Latest release
pip install git+https://github.com/yazirofi/pyllmlib.git
# Development version
git clone https://github.com/yazirofi/pyllmlib.git
cd pyllmlib
pip install -e .
```
**Requirements**:
* Python 3.7+
* `requests` (installed automatically)
---
## 🚀 Quick Start
```python
from pyllmlib import config, generate
# Configure your preferred LLM
config(
provider="openai",
api_key="your-openai-api-key",
model="gpt-4"
)
# Generate text
response = generate("Explain quantum computing in simple terms")
print(response)
```
✅ Same code works with **any provider** — just change the config.
---
## ⚙️ Configuration
### 1. Direct in Code
```python
from pyllmlib import config
# OpenAI
config(provider="openai", api_key="sk-...", model="gpt-4")
# Google Gemini
config(provider="gemini", api_key="AIza...", model="gemini-2.5-flash")
# Mistral
config(provider="mistral", api_key="...", model="mistral-large-latest")
# Groq
config(provider="groq", api_key="gsk_...", model="mixtral-8x7b-32768")
```
### 2. Environment Variables
```bash
LLM_PROVIDER=openai
LLM_API_KEY=sk-your-openai-key
LLM_MODEL=gpt-4
LLM_BASE_URL=https://api.openai.com/v1 # Optional
```
```python
from pyllmlib import config, generate
config() # Loads from env
print(generate("What is LLM?"))
```
---
## 💬 Usage Examples
### Text Generation
```python
from pyllmlib import config, generate
config(provider="gemini", api_key="AIza...", model="gemini-2.5-flash")
print(generate("What is the capital of France?"))
prompt = """
Write a Python function to calculate factorial with error handling and docstring.
"""
print(generate(prompt))
```
### Interactive Chat
```python
from pyllmlib import config, chat, reset_chat
config(provider="gemini", api_key="AIza...", model="gemini-2.5-flash")
while q := input("Ask: "):
print(chat(q))
reset_chat()
```
---
## Style Generated Output Response
```python
from pyllmlib import config, chat, reset_chat, style, generate
config(provider="gemini", api_key="AIza...", model="gemini-2.5-flash")
style(generate("Write the wish in one line")) # use the **style** function instead of **print** function
while q := input("Ask: "):
style(chat(q)) # use the **style** function instead of **print** function
reset_chat()
```
---
## 🌐 Supported Providers
### ✅ Currently Supported
* **OpenAI** — `gpt-4`, `gpt-4-turbo`, `gpt-3.5-turbo`
* **Google Gemini** — `gemini-2.5-flash`, `gemini-1.5-flash`
* **Mistral AI** — `mistral-large-latest`, `mistral-small-latest`
* **Groq** — `mixtral-8x7b-32768`, `llama2-70b-4096`, `gemma-7b-it`
### 🔜 Coming Soon
* Anthropic Claude
* Cohere
* Ollama & LM Studio (local hosting)
* Hugging Face models
---
## 🐛 Troubleshooting
* **Auth Errors** → Check API key format
* **Model Not Found** → Verify model name is correct
---
## 📊 Best Practices
```python
# ✅ Reuse config for multiple prompts
config(provider="openai", api_key="sk-...", model="gpt-4")
for p in prompts:
print(generate(p))
# ❌ Don’t reconfigure on every request
```
💡 **Cost Optimization**: Use `gpt-3.5-turbo` for simple tasks, `gpt-4` for complex ones.
---
## 📚 API Reference
* `config(**kwargs)` → Set provider, API key, model
* `generate(prompt, **kwargs)` → Single text output
* `generate_stream(prompt, **kwargs)` → Streaming output
* `chat(message)` → Conversational interface
* `reset_chat()` → Clear conversation history
---
## 📄 License
Licensed under the **MIT License** – see [LICENSE](LICENSE).
---
## 👨💻 Author
**Shay Yazirofi**
* GitHub: [@yazirofi](https://github.com/yazirofi)
* Email: [yazirofi@gmail.com](mailto:yazirofi@gmail.com)
---
⭐ If you find **pyllmlib** useful, please **star the repo** on GitHub!
📖 More docs & tutorials: [Wiki](https://github.com/yazirofi/pyllmlib/)
---
Raw data
{
"_id": null,
"home_page": "https://github.com/yazirofi/pyllmlib",
"name": "pyllmlib",
"maintainer": null,
"docs_url": null,
"requires_python": ">=3.7",
"maintainer_email": null,
"keywords": "python genai llm api openai gemini mistral groq ai chatbot",
"author": "Shay Yazirofi",
"author_email": "yazirofi@gmail.com",
"download_url": "https://files.pythonhosted.org/packages/c7/ee/32cd9f5888f7629ef8da63a9d17b96f9f4fa40bee0c5f9484b52877432fc/pyllmlib-0.1.1.tar.gz",
"platform": null,
"description": "# \ud83d\udd0c pyllmlib\r\n\r\n[](https://pepy.tech/projects/pyllmlib)\r\n[](https://pypi.org/project/pyllmlib/)\r\n\r\n[](LICENSE)\r\n[](https://github.com/yazirofi/pyllmlib)\r\n\r\n---\r\n\r\n<div style=\"\r\n width:200px; \r\n height:100px; \r\n overflow:hidden; \r\n display:flex; \r\n justify-content:center; \r\n align-items:center; \r\n margin:auto; \r\n position:relative; \r\n top:0; \r\n bottom:0; \r\n left:0; \r\n right:0; \r\n border-radius:10px; \r\n box-shadow:0 4px 8px rgba(0,0,0,0.2);\r\n\">\r\n <img src=\"https://raw.githubusercontent.com/yazirofi/cdn/main/yazirofi.jpg\" \r\n style=\"width:100%; height:100%; object-fit:cover;\" \r\n alt=\"Cropped Image\">\r\n</div>\r\n\r\n---\r\n\r\n**pyllmlib** is a lightweight, provider-agnostic Python package that gives you **one simple interface** to work with multiple Large Language Model (LLM) APIs.\r\n\r\nStop juggling different SDKs and client libraries \u2014 whether it\u2019s **OpenAI GPT**, **Google Gemini**, **Mistral AI**, or **Groq**, you write your code once and switch providers in seconds.\r\n\r\n\ud83c\udfaf **Ideal for** developers who want to:\r\n\r\n* Experiment with different LLMs quickly\r\n* Build multi-provider AI applications\r\n* Avoid vendor lock-in with a consistent API\r\n\r\n---\r\n\r\n## \u2728 Features\r\n\r\n* \ud83d\udd0c **Unified API** \u2014 A single `generate()` function for all providers\r\n* \ud83c\udf10 **Multi-Provider Support** \u2014 OpenAI, Gemini, Mistral, Groq (with more coming soon)\r\n* \ud83e\udde0 **Consistent Message Format** \u2014 Same request style across providers\r\n* \ud83d\udd10 **Flexible Config** \u2014 Use env vars, inline setup, or config files\r\n* \ud83d\udce6 **Minimal Dependencies** \u2014 Only needs `requests`\r\n* \ud83d\udd04 **Quick Provider Switching** \u2014 Change models with one line\r\n* \ud83d\udee1\ufe0f **Automatic Token Handling** \u2014 Prevents overflows & context errors\r\n* \ud83d\udcdc **Role-Based Conversations** \u2014 System, user, assistant messages\r\n* \ud83d\udd27 **Extensible** \u2014 Add your own providers with minimal code\r\n* \ud83d\ude80 **No Vendor Lock-In** \u2014 Swap providers without rewriting logic\r\n\r\n---\r\n\r\n## \ud83d\udce6 Installation\r\n\r\nFrom PyPI (recommended):\r\n\r\n```bash\r\npip install pyllmlib\r\n```\r\n\r\nFrom GitHub:\r\n\r\n```bash\r\n# Latest release\r\npip install git+https://github.com/yazirofi/pyllmlib.git\r\n\r\n# Development version\r\ngit clone https://github.com/yazirofi/pyllmlib.git\r\ncd pyllmlib\r\npip install -e .\r\n```\r\n\r\n**Requirements**:\r\n\r\n* Python 3.7+\r\n* `requests` (installed automatically)\r\n\r\n---\r\n\r\n## \ud83d\ude80 Quick Start\r\n\r\n```python\r\nfrom pyllmlib import config, generate\r\n\r\n# Configure your preferred LLM\r\nconfig(\r\n provider=\"openai\",\r\n api_key=\"your-openai-api-key\",\r\n model=\"gpt-4\"\r\n)\r\n\r\n# Generate text\r\nresponse = generate(\"Explain quantum computing in simple terms\")\r\nprint(response)\r\n```\r\n\r\n\u2705 Same code works with **any provider** \u2014 just change the config.\r\n\r\n---\r\n\r\n## \u2699\ufe0f Configuration\r\n\r\n### 1. Direct in Code\r\n\r\n```python\r\nfrom pyllmlib import config\r\n\r\n# OpenAI\r\nconfig(provider=\"openai\", api_key=\"sk-...\", model=\"gpt-4\")\r\n\r\n# Google Gemini\r\nconfig(provider=\"gemini\", api_key=\"AIza...\", model=\"gemini-2.5-flash\")\r\n\r\n# Mistral\r\nconfig(provider=\"mistral\", api_key=\"...\", model=\"mistral-large-latest\")\r\n\r\n# Groq\r\nconfig(provider=\"groq\", api_key=\"gsk_...\", model=\"mixtral-8x7b-32768\")\r\n```\r\n\r\n### 2. Environment Variables\r\n\r\n```bash\r\nLLM_PROVIDER=openai\r\nLLM_API_KEY=sk-your-openai-key\r\nLLM_MODEL=gpt-4\r\nLLM_BASE_URL=https://api.openai.com/v1 # Optional\r\n```\r\n\r\n```python\r\nfrom pyllmlib import config, generate\r\n\r\nconfig() # Loads from env\r\nprint(generate(\"What is LLM?\"))\r\n```\r\n\r\n---\r\n\r\n## \ud83d\udcac Usage Examples\r\n\r\n### Text Generation\r\n\r\n```python\r\nfrom pyllmlib import config, generate\r\n\r\nconfig(provider=\"gemini\", api_key=\"AIza...\", model=\"gemini-2.5-flash\")\r\n\r\nprint(generate(\"What is the capital of France?\"))\r\n\r\nprompt = \"\"\"\r\nWrite a Python function to calculate factorial with error handling and docstring.\r\n\"\"\"\r\nprint(generate(prompt))\r\n```\r\n\r\n### Interactive Chat\r\n\r\n```python\r\nfrom pyllmlib import config, chat, reset_chat\r\n\r\nconfig(provider=\"gemini\", api_key=\"AIza...\", model=\"gemini-2.5-flash\")\r\n\r\nwhile q := input(\"Ask: \"):\r\n print(chat(q))\r\n\r\nreset_chat()\r\n```\r\n---\r\n\r\n## Style Generated Output Response\r\n```python\r\nfrom pyllmlib import config, chat, reset_chat, style, generate\r\n\r\nconfig(provider=\"gemini\", api_key=\"AIza...\", model=\"gemini-2.5-flash\")\r\n\r\nstyle(generate(\"Write the wish in one line\")) # use the **style** function instead of **print** function\r\n\r\nwhile q := input(\"Ask: \"):\r\n style(chat(q)) # use the **style** function instead of **print** function\r\n \r\nreset_chat()\r\n```\r\n\r\n---\r\n\r\n## \ud83c\udf10 Supported Providers\r\n\r\n### \u2705 Currently Supported\r\n\r\n* **OpenAI** \u2014 `gpt-4`, `gpt-4-turbo`, `gpt-3.5-turbo`\r\n* **Google Gemini** \u2014 `gemini-2.5-flash`, `gemini-1.5-flash`\r\n* **Mistral AI** \u2014 `mistral-large-latest`, `mistral-small-latest`\r\n* **Groq** \u2014 `mixtral-8x7b-32768`, `llama2-70b-4096`, `gemma-7b-it`\r\n\r\n### \ud83d\udd1c Coming Soon\r\n\r\n* Anthropic Claude\r\n* Cohere\r\n* Ollama & LM Studio (local hosting)\r\n* Hugging Face models\r\n\r\n---\r\n\r\n## \ud83d\udc1b Troubleshooting\r\n\r\n* **Auth Errors** \u2192 Check API key format\r\n* **Model Not Found** \u2192 Verify model name is correct\r\n\r\n---\r\n\r\n## \ud83d\udcca Best Practices\r\n\r\n```python\r\n# \u2705 Reuse config for multiple prompts\r\nconfig(provider=\"openai\", api_key=\"sk-...\", model=\"gpt-4\")\r\nfor p in prompts:\r\n print(generate(p))\r\n\r\n# \u274c Don\u2019t reconfigure on every request\r\n```\r\n\r\n\ud83d\udca1 **Cost Optimization**: Use `gpt-3.5-turbo` for simple tasks, `gpt-4` for complex ones.\r\n\r\n---\r\n\r\n## \ud83d\udcda API Reference\r\n\r\n* `config(**kwargs)` \u2192 Set provider, API key, model\r\n* `generate(prompt, **kwargs)` \u2192 Single text output\r\n* `generate_stream(prompt, **kwargs)` \u2192 Streaming output\r\n* `chat(message)` \u2192 Conversational interface\r\n* `reset_chat()` \u2192 Clear conversation history\r\n\r\n---\r\n\r\n\r\n## \ud83d\udcc4 License\r\n\r\nLicensed under the **MIT License** \u2013 see [LICENSE](LICENSE).\r\n\r\n---\r\n\r\n## \ud83d\udc68\ud83d\udcbb Author\r\n\r\n**Shay Yazirofi**\r\n\r\n* GitHub: [@yazirofi](https://github.com/yazirofi)\r\n* Email: [yazirofi@gmail.com](mailto:yazirofi@gmail.com)\r\n\r\n---\r\n\r\n\u2b50 If you find **pyllmlib** useful, please **star the repo** on GitHub!\r\n\ud83d\udcd6 More docs & tutorials: [Wiki](https://github.com/yazirofi/pyllmlib/)\r\n\r\n---\r\n\r\n",
"bugtrack_url": null,
"license": "MIT",
"summary": "A Simple Unified LLM API interface for OpenAI, Gemini, Mistral, Groq and more.",
"version": "0.1.1",
"project_urls": {
"Documentation": "https://github.com/yazirofi/pyllmlib#readme",
"Homepage": "https://github.com/yazirofi/pyllmlib",
"Source Code": "https://github.com/yazirofi/pyllmlib"
},
"split_keywords": [
"python",
"genai",
"llm",
"api",
"openai",
"gemini",
"mistral",
"groq",
"ai",
"chatbot"
],
"urls": [
{
"comment_text": null,
"digests": {
"blake2b_256": "c4d2672e46de9d636af6cb555740b8a6440c25e1f3adca69a9594833ccd7564e",
"md5": "b075dc195161e260bc701f3c03dd012a",
"sha256": "490f47063e9da6711438c5cadd464b36d40ad140c652401f33d957f30e72d745"
},
"downloads": -1,
"filename": "pyllmlib-0.1.1-py3-none-any.whl",
"has_sig": false,
"md5_digest": "b075dc195161e260bc701f3c03dd012a",
"packagetype": "bdist_wheel",
"python_version": "py3",
"requires_python": ">=3.7",
"size": 10566,
"upload_time": "2025-08-22T18:46:14",
"upload_time_iso_8601": "2025-08-22T18:46:14.005626Z",
"url": "https://files.pythonhosted.org/packages/c4/d2/672e46de9d636af6cb555740b8a6440c25e1f3adca69a9594833ccd7564e/pyllmlib-0.1.1-py3-none-any.whl",
"yanked": false,
"yanked_reason": null
},
{
"comment_text": null,
"digests": {
"blake2b_256": "c7ee32cd9f5888f7629ef8da63a9d17b96f9f4fa40bee0c5f9484b52877432fc",
"md5": "906919a2bee5918bb0ea73246c2383af",
"sha256": "ffa0821d5184a9ac842c7118af88127892369bf112ada2420ea66060193f6cfc"
},
"downloads": -1,
"filename": "pyllmlib-0.1.1.tar.gz",
"has_sig": false,
"md5_digest": "906919a2bee5918bb0ea73246c2383af",
"packagetype": "sdist",
"python_version": "source",
"requires_python": ">=3.7",
"size": 10934,
"upload_time": "2025-08-22T18:46:15",
"upload_time_iso_8601": "2025-08-22T18:46:15.082365Z",
"url": "https://files.pythonhosted.org/packages/c7/ee/32cd9f5888f7629ef8da63a9d17b96f9f4fa40bee0c5f9484b52877432fc/pyllmlib-0.1.1.tar.gz",
"yanked": false,
"yanked_reason": null
}
],
"upload_time": "2025-08-22 18:46:15",
"github": true,
"gitlab": false,
"bitbucket": false,
"codeberg": false,
"github_user": "yazirofi",
"github_project": "pyllmlib",
"travis_ci": false,
"coveralls": false,
"github_actions": false,
"lcname": "pyllmlib"
}
JSON
