Metadata-Version: 2.4
Name: runlatent
Version: 0.0.1
Summary: Latent: runtime reliability monitoring for LLM answers. One line to instrument OpenAI, Anthropic and Gemini calls. Placeholder release; the first real release follows.
Author-email: Latent <hello@runlatent.ai>
License: Proprietary
Project-URL: Homepage, https://runlatent.ai
Project-URL: Documentation, https://runlatent.ai/docs
Classifier: Development Status :: 1 - Planning
Classifier: Programming Language :: Python :: 3
Requires-Python: >=3.10
Description-Content-Type: text/markdown

# runlatent

Latent reads every finished LLM answer and tells you which ones are likely unsupported by the source the model was given, so your team reviews the few that matter.

This is a name reservation. The first working release ships shortly:

    pip install runlatent

    import runlatent
    runlatent.auto_instrument()

with `LATENT_API_URL` and `LATENT_API_KEY` set. Documentation: https://runlatent.ai/docs
