Metadata-Version: 2.1
Name: unreasonable-llama
Version: 0.2.5
Summary: HTTP API bindings for llama.cpp built-in server
Home-page: https://github.com/SteelPh0enix/unreasonable-llama/
License: MIT
Keywords: llama.cpp,llama,llm,api,bindings
Author: SteelPh0enix
Author-email: wojciech_olech@hotmail.com
Requires-Python: >=3.12,<4.0
Classifier: License :: OSI Approved :: MIT License
Classifier: Programming Language :: Python :: 3
Classifier: Programming Language :: Python :: 3.12
Requires-Dist: httpx (>=0.27.0,<0.28.0)
Project-URL: Repository, https://github.com/SteelPh0enix/unreasonable-llama/
Description-Content-Type: text/markdown

# unreasonable-llama

[![Check code formatting and validity](https://github.com/SteelPh0enix/unreasonable-llama/actions/workflows/check-code.yml/badge.svg?branch=master)](https://github.com/SteelPh0enix/unreasonable-llama/actions/workflows/check-code.yml)

(Yet another) Python API for [llama.cpp server](https://github.com/ggerganov/llama.cpp/tree/master/examples/server)

For now, i'm targeting minimal support necessary for `/completion` and `/health` endpoint.
Maybe i'll extend this lib in the future.

## Requirements

`unreasonable-llama` has a single requirement - `httpx` library, see `pyproject.toml` for details.

Requirements for `llama.cpp` scripts can be found in `requirements/` directory of theirs repository.

## Usage

See [test file](./tests/__main__.py) for example usage.

