MerchantryTidbits

evals-testing / library

AIConfig

Symptoms: hardcoded prompts per provider; config drift across environments.

Use it when

  • config-driven prompts across model providers
  • version prompt and model settings outside code

What it solves

multi-model-routingmessy-json-yaml

Not the fit when

  • not an inference server

Install

pip install python-aiconfig

Invoke

Store prompts and model params in .aiconfig.json and load across providers.

Alternatives

No reviewed alternatives recorded yet.