Skip to content

Latest commit

 

History

History
77 lines (43 loc) · 2.06 KB

File metadata and controls

77 lines (43 loc) · 2.06 KB

Launch Kit

One-line pitch

ai-stability is a Python CLI that runs the same prompt multiple times, scores LLM output consistency, and shows where responses vary.

Short post

I shipped ai-stability, a CLI-first tool for measuring how stable an LLM is on the same prompt.

Useful for prompt testing, model comparison, and debugging flaky AI behavior.

X / Twitter post

Shipped ai-stability today.

A small Python CLI that runs the same prompt multiple times, compares the outputs, scores consistency, and shows diffs.

Useful for prompt testing and debugging flaky LLM behavior.

pipx install ai-stability

GitHub: https://github.com/BuildWithAbid/ai-stability PyPI: https://pypi.org/project/ai-stability/

LinkedIn post

I built ai-stability, a CLI-first Python tool for checking how stable an LLM is on repeated runs of the same prompt.

The tool:

  • runs the same prompt multiple times
  • compares response similarity
  • assigns a simple stability score
  • highlights where outputs drift
  • saves a local JSON artifact for replay and debugging

This is useful when you are testing prompts, comparing model behavior, or trying to understand whether “randomness” is minor wording noise or real behavior variance.

Install:

pipx install ai-stability

Project:

Show HN

Title

Show HN: ai-stability – CLI tool to measure LLM output stability

Body

I built a small Python CLI called ai-stability.

It runs the same prompt multiple times against the same model, compares the outputs, computes a simple stability score, shows diffs between runs, and saves a local JSON artifact.

It is aimed at developers testing prompts, comparing model behavior, or debugging unstable AI output.

Install:

pipx install ai-stability

Repo:

https://github.com/BuildWithAbid/ai-stability

PyPI:

https://pypi.org/project/ai-stability/