Moose can think on
your machine.
Want AI that never leaves your desk? Moose runs a real Gemma model right on your laptop, entirely offline - unlimited, private, and free. And it's a feature, not a requirement: go managed or bring your own OpenRouter key whenever you want frontier models doing the thinking.
Runs offline. Works on Windows & macOS.
A model that fits the laptop you already have.
Gemma comes in three sizes. Start small on an older machine, or load the big one on a workstation for the deepest analysis. The model downloads once, then runs fully offline - and you can switch any time.
The everyday default. Strong briefs and insights with no perceptible wait. Most people stay here.
Local means local. Your work stays on your disk.
When you run the local model, the AI lives next to your files: prompts, drafts, and context never leave the machine. The cloud only comes into it through something you set up yourself: a managed plan, an OpenRouter key you add, or a CMS you connect to publish.
Things go out when you choose cloud: a managed plan where we run the frontier models for you, your own OpenRouter key, or a CMS you connect to publish. Each one is opt-in.
Want a frontier model in the loop? Plug in your key.
When a job calls for ChatGPT, Claude, or Gemini, connect your OpenRouter key and Moose routes that one task through it. You pay OpenRouter at the provider's rate - never a markup to us. Prefer zero setup? The managed plan runs frontier models for you, no key needed. Either way, the local model is always there, free.
ChatGPT
Claude
Gemini+ moreLocal AI Privacy and Models FAQs
How does Hi, Moose protect my data?
When you use the local model, your prompts, context, drafts, brand voice, memory, library, and connected Search Console data stay on your machine. The model runs alongside your files, and its downloaded model can operate fully offline. Cloud services are used only when you explicitly choose them.
What information leaves my computer when I use local AI?
Nothing leaves your computer for local-model tasks. Information is sent out only when you opt into a cloud action, such as using a managed plan, routing a task through an OpenRouter key you added, or publishing to a CMS you connected.
Can I use Hi, Moose AI without per-query costs?
Yes. The local Gemma model has no per-token or per-query bill. You can chat, create briefs, and draft without local usage limits. If you choose to route specific tasks through OpenRouter, those tasks are billed by OpenRouter at the provider's rate.
Which local AI model sizes are available?
Hi, Moose offers Gemma in three local sizes: Light Gemma 2B, Balanced Gemma 4B, and Max Gemma 26B. Light is designed for older machines with 8 GB RAM, Balanced is the everyday option with 16 GB RAM, and Max is for deeper analysis on workstations with 32 GB RAM. You can switch sizes whenever you need.
Can I connect ChatGPT, Claude, or Gemini to Hi, Moose?
Yes. Add your own OpenRouter key to route selected tasks to frontier models such as ChatGPT, Claude, or Gemini. Your key is stored on your machine, not on Hi, Moose servers. You can also choose a managed plan for frontier-model access without managing a key, while keeping the local model available for free.
Your laptop can do
the thinking.
Free to download, unlimited local use, private by default. Add your OpenRouter key or a managed plan whenever you want frontier power.