Backengine is a tool that evolved from an experiment at my startup for which we wanted to pass structured data to an LLM completion API and receive structured data back.<p>We thought we'd release it as a standalone tool for others who might also find it useful.<p>Backengine lets you create hosted API endpoints for which the endpoint logic is described in natural language as an LLM prompt. The endpoints accept structured requests and return structured responses.<p>Why is this useful?<p>Hosted - no additional infrastructure to maintain.<p>Prompt engineering workspace - you can edit, test and tweak endpoint prompts without touching any code.<p>Instant deployment - created endpoints should be immediately available, and changes to your endpoints are deployed in one click.<p>This is a first and very early release - we'd love for HN to try out Backengine and hear what you think! There's a playground where you can try it out - we're very interested to hear what use-cases people come up with.
Let's get this clear, the endpoint responses are hallucinated by an LLM, right? So this is mostly a tool to experiment with endpoint design in a non-committal way? Or am I missing something fundamental here?