> For the complete documentation index, see [llms.txt](https://ultrasafe.gitbook.io/docs/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://ultrasafe.gitbook.io/docs/the-platform.md).

# The Platform

UltraSafe AI offers versatile access to our Large Language Models through three key options:

1\. **Direct API Usage**: Access our latest models via API endpoints on a pay-as-you-go basis through our dedicated platform.   

2\. **Cloud Integration**: Easily incorporate UltraSafe AI models into your cloud infrastructure.

3\. **On-Premises Hosting** : Self-deploy our open-weight models on your own servers. These models are distributed under the Apache 2.0 License and are available for download from Hugging Face or our documentation.

<div align="left"><figure><img src="https://lh7-rt.googleusercontent.com/docsz/AD_4nXfvfK7ll-41i-y0rUHo8EOdhEt9AyxIYIKrvAO-8caN44okI9IDPImwllBeGu4KndKKu3p0bmHhzbopvBg23ep5Hw5u4KrH-4UFUD1zXf9zksj9RrYVKZzqmo85lv_TszLd8_gz_K9ExN4rYhgxxU2TWQJQ?key=rJldTYnqSOCJnhAWBD4HIg" alt=""><figcaption></figcaption></figure></div>

### **Getting Started with API Access on The Platform**

First, activate payments on your account to unlock your API keys on The Platform. Our Quickstart guide is available to help you make your first UltraSafe API request.

Key features our models offer:

1. **Text Completion**
2. **Embeddings**
3. **Function Calls**
4. **JSON Mode**
5. **Guardrailing**

### Cloud-Based Deployments

For a detailed overview of how to deploy and utilize UltraSafe AI models in the cloud, please refer to our Cloud Deployment section.

### Utilizing Raw Model Weights

Raw model weights can be leveraged in various contexts:

1. **Research Applications**: Visit our reference implementation repository for research-focused resources and examples.
2. **Self-Deployment**: Deploy on cloud or on-premises using TensorRT-LLM or vLLM. Detailed instructions can be found in our Deployment section.
3. **Local Deployment**: For deployment on consumer-grade hardware, explore the llama.cpp project or Ollama for suitable solutions.