textgen

Why we included this project

Most people running local LLMs end up gluing together a model loader, a chat interface, and an API server. textgen skips that assembly: it's a portable desktop app for Windows, Linux, and macOS, and once you drop a GGUF model into a folder you have a working chat window in minutes. The same app also speaks the OpenAI and Anthropic API formats, so existing applications can point at it as a fully offline replacement for those cloud services. It handles image input and file attachments, supports tool-calling, and lets you switch between inference backends like llama.cpp and Transformers without restarting. If privacy, no telemetry, and predictable local costs matter to you, this is a practical way to serve models without standing up a separate server stack.