Voxint bundled local-LLM service — third-party model attribution
=================================================================

This image bundles a vendored large-language-model weight file so that a
single-operator Voxint install can run transcript enhancement and run-asset
summary/entity extraction with no external API key.

Model
-----
  Qwen3-4B-Instruct-2507
  Copyright (c) Alibaba Cloud / Qwen Team.
  Licensed under the Apache License, Version 2.0.
  Upstream model card: https://huggingface.co/Qwen/Qwen3-4B-Instruct-2507

Vendored GGUF quantization
--------------------------
  Qwen3-4B-Instruct-2507-Q5_K_M.gguf
  Re-quantized to GGUF (Q5_K_M) by the Unsloth project.
  Licensed under the Apache License, Version 2.0.
  Source: https://huggingface.co/unsloth/Qwen3-4B-Instruct-2507-GGUF
  Revision: a06e946bb6b655725eafa393f4a9745d460374c9
  sha256:  5bde5e9d883622acb02bf77fe7dcbc56a8b9a9ad4be78a72ca23a532658b4ecb

Serving engine
--------------
  llama.cpp server (ggml-org/llama.cpp), also Apache-2.0 / MIT licensed.
  https://github.com/ggml-org/llama.cpp

A full copy of the Apache License, Version 2.0 is available at
https://www.apache.org/licenses/LICENSE-2.0. Voxint itself is Apache-2.0.
