Introduction

What is the platform?

Generative AI is a powerful tool that is rapidly being integrated into new and old software applications because of its ability to derive impressively accurate answers based on the data that is made available to it. The vulnerability with generative AI though is that these answers are reached probabilistically, and therefore there is always the chance that systems built on top of it can hallucinate, meaning product outputs that sound plausible but are actually false.

While this is not overly problematic for some use cases, for many it is an unacceptable outcome. The Jaxon platform serves to meet the needs of these users by providing a suite of guardrails that can sit on top of an LLM-based system, ensuring that the outputs stay within desired constraints.

How do I use it?

The platform provides both a web-based administration interface and a programmatic API for integration with your applications.

Web Interface

The platform includes a comprehensive web UI that allows you to: - Configure applications and guardrails - Test guardrails interactively using the Playground - Review and correct guardrail outputs with Human Review - Monitor system activity and performance

Access the web interface by navigating to your platform deployment URL.

Programmatic Integration

For production use, the platform provides client libraries that integrate with your applications. See the Client API documentation for detailed information about: - Installing and configuring the Python client - Connecting via HTTP/WebSocket or direct Kafka - Sending requests and receiving responses - Authentication and security

Who is it for?

The Jaxon platform is for anyone building with LLMs who wants more control over their outputs, as well as those building applications that require precision analysis, review, and monitoring of textual data.