kuafuai/devopsgpt

DevOpsGPT is a multi-agent system that takes a written software requirement in plain language and tries to turn it into working, deployed code. It connects a large language model to DevOps tooling so that the whole pipeline from requirement clarification through code generation, continuous integration, and deployment can run with less manual intervention. The general workflow starts when you describe what you want built. DevOpsGPT asks clarifying questions to refine the requirement, then generates interface documentation, writes pseudocode based on any existing project code, and produces working code from that pseudocode. From there it hooks into DevOps tools for automated testing and deployment to a target environment. The README notes two current limitations: generated requirements and interface documentation may not be precise enough in complex scenarios, and automatic understanding of existing project code is still incomplete, with a future version planned to address it. The tool works with any programming language and is intended to extend existing projects, not just generate greenfield code. An enterprise edition adds deeper existing-project analysis, support for language models beyond GPT, private deployment, and connections to more DevOps platforms. The open-source version requires a GPT API token and runs locally via a Python script or a Docker container. Once running, you interact through a browser interface at a local address. Setup for the self-hosted version takes a few steps: download the release or clone the repo, copy the template configuration file, add your API token, and run a startup script. Docker users mount a workspace directory and the config file, then start the container. Generated code lands in a local workspace folder. A cloud-hosted version is available at kuafuai.net for those who prefer not to self-host. The project is experimental and comes with a disclaimer that the developers accept no liability for issues arising from its use.