TextGen
Open-source browser interface for running language models offline on your own computer, including an API and LoRA training.
What it does
Text Generation WebUI, now listed under the name TextGen, is a browser interface for running language models on your own computer. Backends include llama.cpp, ik_llama.cpp, Transformers, ExLlamaV3 and TensorRT-LLM, and you can switch between them without restarting. You chat in instruct or chat mode, attach images, PDFs and Word files, and branch conversations. On top of that come a notebook for free-form text, LoRA fine-tuning, image generation via diffusers models, tool calling with MCP servers and an API compatible with OpenAI and Anthropic. Portable builds are available for Windows, macOS and Linux, plus Docker images.
Who it suits
Developers and technically skilled teams who want to test or compare open models, or connect them to their own applications via API, without sending data to a cloud service.
Data protection for businesses
The software runs locally and, according to the project, has no telemetry, external resources or update requests. Inputs stay on your computer or server and are not used for training; data lives wherever you run the software. There is no data processing agreement, because no vendor processes your data. The project offers no business plans.
Limits
Performance depends on your hardware; larger models need a powerful graphics card. Setup and model selection require technical knowledge. Help comes only from the community, and the AGPL-3.0 licence matters if you redistribute it or offer it over a network.