Lantern - Ship production quality code with AI

Think of Lantern as AI's Guide.

We've all heard "AI is great at boiler-plate", but the reality is it kind of sucks at it. LLMs may produce code that compiles, but it is often "average" github side-project code that as it grows in complexity starts to fall apart. The conventions and style aren't clear up front which leads to inconsistencies. As the code-base grows we end up with:

• Confused AI agents
• Declining code quality
• Frustrated, less efficent developers

Lantern looks to aleviating these issues by providing developers with the tools to build on top of high quality projec

1. Formal Specification Language - Lantern Script is a simple systems description langague that lookes like plaine english that serves as your projects formal specifications, documentation, and BDD verificaiton framework.

2. Templates

3. BDD / Verification

Getting Started

Create a new project from a lantern template

lantern new <project_name> <template>

For a full list of available project templates use lantern template list

For a full list of template commands run lantern template --help

Building Lantern

First you'll need to install rust, then run: cargo install

Goals

Lantern Script

Lantern aims to be a new language paradim, but inherits syntax / features from other languages.

• Superset of Gherkin. This allows you to write high level Cucumber test descriptions.
• Python style docstrings

Templates

Lantern provides a template repository for creating project scaffolding that provide a basis for new projects that provide a high quality code structure for AI to build on top of.

Automated Verification