When working on DSL grammars, I kept feeling that the workflow inside the editor was more fragmented than it should be.
I wanted a flow where I could understand the current grammar, draft a scaffold, generate sample DSL, and validate it without constantly jumping between disconnected tools. Most existing AI tools felt too generic for that.
So I built DSLForge, a VS Code extension for Langium, ANTLR4, and Xtext workspaces.
Its current features are simple but focused:
- explain the current grammar
- generate DSL scaffold proposals
- generate sample DSL
- validate the current grammar
The key design decision is that Validate Current Grammar is intentionally non-AI. Instead of faking validation, DSLForge follows the project’s real validation route through configured commands, package scripts, Gradle wrapper tasks, or Maven wrapper goals.
I wanted the extension to be useful for actual DSL authoring work, not just as a generic text generator.
If you work with language tooling, grammar authoring, or DSL-heavy projects, I’d be interested in your feedback.

Top comments (0)