Observed arrival · 2026-08-29
digline: Regression Tests for LLM Applications
An open-source command-line tool that compares LLM application runs against an approved JSON baseline stored in your repository.
Field notes
The tool treats an LLM test baseline as a JSON artifact inside the application repository rather than as a hosted dashboard record. Its sample output separates regressions by named case and check, including a latency score falling from 0.816327 to 0.650407 and a regex check changing from passing to failing. Installation is shown through pip, with Python 3.12+ required, and the project links to public GitHub, PyPI, reference, and example materials.
Observed signals
Read the marks
Editorial observations of this landing page, not a rating.
One card from the complete issue