Questions? Answers.

Is grAIm free?

Yes, grAIm is free during beta. You bring your own API keys for LLM access — you only pay your LLM provider directly for the API usage (and many offer free tiers). We'll announce pricing before beta concludes. grAIm will remain free to use with a generous free tier — extended usage, sync, and marketplace access will require a paid account (starting around $10/month) after that.

Is my data private?

All memory is on your device. Conversations go direct to your LLM provider — we never see them. Your memory is encrypted on-device, accessible only through your account login, and exports can be encrypted or plain text — your choice. During beta, anonymised diagnostic data is collected to find and fix bugs (required for beta participation). See the full privacy architecture.

What LLMs does grAIm support?

Claude, ChatGPT, Gemini, DeepSeek, Z AI, OpenRouter (200+ models), and any OpenAI-compatible API. You can even use different providers for different jobs — one model for responses, another for memory extraction, a third for judgement. Don't see yours? Email us or join our Discord to suggest it.

Why don't you support local models?

Models that can run on a mobile device today are too limited to power a useful experience with grAIm. They're slow, they drain the battery, and the ones small enough to fit struggle with anything beyond simple replies. If you do want to run a local model, you can host one on your own server and connect grAIm to it through any OpenAI-compatible API — that gives you local control without the mobile trade-offs. But running a model directly on your phone just isn't practical yet — the moment on-device models are ready, we'll be the first to bring them back.

What devices are supported?

Android now. iOS and desktop (Windows, Mac, Linux) coming later. The beta is Android-first — sign up and you'll get access as soon as your platform is ready.

What happens after beta?

Your memories and assistant identity carry over to the release version. Diagnostic data is deleted when beta ends. We'll announce pricing before beta concludes — but your data is yours and always will be: you can export your memory and assistant identity as a JSON file at any time, so if you ever want to take your assistant elsewhere, your data goes with you.

How much does the LLM usage cost me?

Usually far less than a subscription. You pay your provider per token, and grAIm shows the running cost of every conversation right in the app — long chats with budget models often cost fractions of a cent. After beta, extended usage will require a paid account — the free tier will be enough to properly try the app, but heavy daily use and premium features like sync and the marketplace will be paid.

Does grAIm work offline?

Voice (speech-to-text and text-to-speech) runs on-device by default, online or not — so dictation and spoken replies work with no signal at all. Cloud voice providers are available if you want them. Your chosen LLM runs in the cloud, so the assistant needs a connection to reply.

Is my API key safe?

Keys are stored in your device's secure storage and used only to call your provider directly from your phone. They are never sent to grAIm servers, never logged, and never leave your device except in requests to the provider they belong to.

Can I see and edit what my assistant remembers?

Yes — that's the point. The Memory Browser lists every memory with its tags, priority, and token count. You can edit, share, or delete any entry. Nothing your assistant knows is hidden from you.

Will the animated avatar drain my battery?

No — .gra avatars use lightweight state-aware animation (idle, listening, thinking, speaking) designed specifically to be efficient on phones. And if you prefer, a static image uses much less battery than an animated avatar.