LanzerHow well do agents write your DSL?

Campaigns of programs, written by coding agents, checked by your Langium language's own parser, validator and runtime.

Evaluate agents and skills

Run a suite several times, get a pass rate per campaign and the stage each failure reached, and compare two reports.

Every sample that passes parses, validates, uses what it was asked to and prints what it should — examples, fixtures, training data.

Run blocks execute the program; negative files must be rejected with an exact diagnostic.

Claude Code, Codex, Gemini CLI — anything that speaks the Agent Client Protocol.