Metadata-Version: 2.4
Name: instar-harness
Version: 0.0.0
Summary: Placeholder for Instar (import/CLI: instar), an open-source LLM measurement harness. Not yet released.
Author: Purple Blossom AI
License: Apache-2.0
Project-URL: Homepage, https://github.com/PurpleBlossomAI/instar
Project-URL: Source, https://github.com/PurpleBlossomAI/instar
Keywords: llm,evaluation,measurement,harness,routing
Classifier: Development Status :: 1 - Planning
Classifier: Intended Audience :: Developers
Classifier: Programming Language :: Python :: 3
Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
Requires-Python: >=3.10
Description-Content-Type: text/markdown

# Instar

**Name placeholder — not yet released.**

Instar is an open-source LLM measurement harness (Apache-2.0). It runs your own
workloads through candidate models, providers, and routing policies and produces
defensible cost / quality / latency numbers.

This `0.0.0` release reserves the **`instar-harness`** distribution name on PyPI
(the bare `instar` name is disallowed by PyPI's name policy). The import package
and CLI remain `instar` — you will `pip install instar-harness` and then use
`import instar` / the `instar` command. The first usable release will be `0.1.0`.
Follow development at https://github.com/PurpleBlossomAI/instar
