llm-distillery is an open-source Python package for capturing LLM traffic through a drop-in proxy, curating the resulting interactions into datasets, and distilling smaller models. It is intended for developers and ML engineers optimizing or replacing frontier-model calls.
llm-distillery sits in PulseGate's Fine-tuning & training category. It focuses on replacing expensive frontier-model usage with smaller models trained on curated production traffic. It is built as an open-source project for machine learning engineers and developers building LLM applications. The project is open source (Apache-2.0). It ships for the command line and API, and it can be self-hosted.
llm-distillery first shipped in 2026. Key capabilities include LLM traffic proxy, dataset curation, and model distillation. It exposes integrations via a public API.
Summary written by a language model from the project’s public pages.
What PulseGate has recorded for this listing
Closest matches by what these projects do