About Mmojo Server
#MeWriting Mmojo Server is software you run on your PC or laptop to give you a private, local large language model (LLM) with which you can interact using your web browser. Think ChatGPT, but private with no monthly bill. Your cues, prompts, completions, and answers do not leave your computer and are not logged anywhere on your computer.
Mmojo Server is built from llama.cpp, with an OpenAI-compatible API for developer applications.
User Interface (UI)
Mmojo Server offers the humane Mmojo Complete UI. It is a completion-style UI where you and Mmojo Server collaborate in one shared space. You write. Mmojo Server completes, because “completing” is what LLMs actually do.
Mmojo Complete is my original creation. It is my attempt to strip away the illusion and limitations of chat so that you can use LLMs to their realistic potential. You don’t have to pretend to cosplay with a robot.
Mmojo Complete works with your web browser’s bookmarking system to help you create simple yet powerful AI applications. Tell stories, summarize podcast transcripts, write LinkedIn posts.
Mmojo Chat is a traditional chat-style UI. Saved chats are stored in your browser. Mmojo Chat is a branded version of the chat UI that comes with llama.cpp.
Mmojo Complete and Mmojo Chat are private. None of your data leaves your computer. Both UIs remind you of that prominently.
Works with .gguf Models
Mmojo Server is based on the (near) latest llama.cpp. It works with .gguf models.
OpenAI Compatible API
Developers will appreciate Mmojo Server’s OpenAI compatible API. This lets you develop your AI application on your laptop, without wasting money on cloud tokens. You might even decide that local and private makes sense for your clients or customers too!
Open Source Upgrade Path
Start with the free Mmojo Server, get familiar with the UI and workflow, and connect your AI application. Make sure it works. Then use my GitHub repository to build your own Mmojo Server tuned to your computer’s CPU and GPU hardware. Or have me build it for you or teach you how to build it.
I have chosen the open source path so that my claims of private and local can be evaluated and verified, and so that you can safely ditch the cloud for your LLM needs. No paying for tokens. No surrendering your privacy and dignity. Verifiable trust.
Mmojo Server is and will remain the easiest way for developers to approach and build llama.cpp. This objective and the resulting process keep the Mmojo Server you use easy and focused.
Signed, Brad Hutchings
The use of “I”, “me”, and “my” above is intentional, not informal. I want you to know that a human being created this product, documented this product, tests it, improves it, communicates with real users to get feedback, etc. This isn’t someone’s “AI agent”. I take responsibility (and credit) for what I do. I give credit (and responsibility) to people deserving of it. I hope you expect this from your software developers and solution providers.
Drop me an email if you would like to talk more about what private, local LLMs and Mmojo Server can do for you.
-Brad
Brad Hutchings
brad@BradHutchings.com