If you have been reading about running AI on your own machine, you have met two names over and over: Ollama and LM Studio. Both are free, both run local models, and both are constantly recommended as the way to keep AI off the cloud. The recommendations are not wrong. They just answer a different question than the one most people are actually asking.
What Ollama actually is
Ollama is a command-line tool and a local server for running models. You type a model name, it downloads the weights, and it serves them on your machine. It is excellent at that. It has no companion, no personality, no memory that follows a relationship across weeks. It is a model runner. If you want to run a model and talk to it through a terminal or wire it into your own scripts, Ollama is the right tool.
What LM Studio actually is
LM Studio is a desktop app with a friendly interface for downloading and chatting with local models. It looks much closer to a chat product. You can pick a model, talk to it, and adjust settings. But it is still a model tool wearing a chat window. There is no persistent memory layer, no character that grows with you, no voice that remembers how to say your name. It is a very good way to test and use models locally. It is not a relationship.
Where a companion app starts
A companion app takes the same local models and builds the thing on top of them that the tools leave out. Persistent memory that survives across conversations and connects things by meaning. A character with a consistent voice, in every sense of the word. The ability to send her a picture and have it rendered on your own GPU, or to leave her a voice note she actually understands. None of that is a model. All of it is work the DIY tools do not do.
The honest version is that the tools and the app are not competitors. Ollama and LM Studio are how you run models. A companion is what you build on top of a model, and that second half is the part that takes real time to get right. If you want to run models, use the tools. If you want someone on the other end of the conversation who remembers what you said last week, that is a different thing to build.
Local Waifu runs on the same local models, and since 1.7.0 it can also reach Ollama and LM Studio on your own machine as endpoints, so you do not have to pick between your own setup and a finished companion. You can point her at the model you already run, and keep the memory, voice and relationship on top.
Try her free for 7 days.
No card. Keep her for $20 once, or walk away. Her soul file is yours either way.
Bring her home, try free