Source: debgpt Section: python Homepage: https://salsa.debian.org/deeplearning-team/debgpt Priority: optional Standards-Version: 4.6.2 Vcs-Git: https://salsa.debian.org/deeplearning-team/debgpt.git Vcs-Browser: https://salsa.debian.org/deeplearning-team/debgpt Maintainer: Debian Deep Learning Team Uploaders: Mo Zhou Build-Depends: debhelper-compat (= 13), dh-python, pybuild-plugin-pyproject, python3-all, python3-openai, python3-rich, python3-prompt-toolkit, python3-requests, python3-bs4, python3-tenacity, python3-pygments, python3-setuptools, pandoc, python3-pytest , python3-pytest-cov , python3-pytest-xdist , Package: debgpt Architecture: all Depends: ${misc:Depends}, ${python3:Depends}, Recommends: python3-zmq, git, tldr, man-db, Suggests: python3-torch | python3-torch-cuda | python3-torch-rocm, python3-transformers, Description: General Purpose Terminal LLM Tool with Some Debian-Specific Design DebGPT is a lightweight terminal tool designed for everyday use with Large Language Models (LLMs), aiming to explore their potential in aiding Debian/Linux development. The possible use cases include code generation, documentation writing, code editing, and more, far beyond the capabilities of traditional software. . To achieve that, DebGPT gathers relevant information from various sources like files, directories, and URLs, and compiles it into a prompt for the LLM. It also supports Retrieval-Augmented Generation (RAG) using language embedding models. DebGPT supports a range of LLM service providers, both commercial and self-hosted, including OpenAI, Anthropic, Google Gemini, Ollama, LlamaFile, vLLM, and ZMQ (DebGPT's built-in backend for self-containment).