oobaboogatextgen
Open-source desktop app for local LLMs. Text, vision, tool-calling, OpenAI/Anthropic-compatible API. 100% private.
TL;DR · 30-second scan
textgen (Python) — Open-source desktop app for local LLMs. Text, vision, tool-calling, OpenAI/Anthropic-compatible API. 100% private.
You want fine control over samplers, loaders and prompt formats while running models on your own GPU.
AI & ML · Self-Hosted
A web interface for running text-generation models locally, supporting many loaders — llama.cpp, Transformers, ExLlama — behind a single UI with an extension system. Its reason to exist is control: sampler parameters, prompt formats, LoRA loading and generation internals are all exposed rather than hidden, which is what you want when you are studying model behaviour rather than just chatting. AGPL-3.0, Python. The tradeoff is that the exposed complexity is the product. Newcomers find the interface bewildering next to Open WebUI, dependency installation across GPU vendors is fiddly, and the project moves in bursts. For experimentation it is still the deepest option.
You want fine control over samplers, loaders and prompt formats while running models on your own GPU.
You want a clean chat interface for daily use — Open WebUI is the better fit.
Add this badge to your README to show your project is curated on StackPicks. Free, lightweight (180×28 SVG), and gives your visitors a one-click way to see honest take + alternatives.
[](https://stackpicks.dev/repo/oobabooga-textgen)
<a href="https://stackpicks.dev/repo/oobabooga-textgen"><img src="https://stackpicks.dev/api/badge/oobabooga-textgen" alt="Featured on StackPicks" width="180" height="28" /></a>
Are you the maintainer of oobabooga/textgen? Add the badge and we'll feature your project in the weekly curator newsletter.