Frequently Asked Questions

Is my data private?

Yes. Memo is local-first: all data stays on your device by default. Conversations, memories, and configuration never leave your machine unless you explicitly enable Cloud Sync, which is end-to-end encrypted.

Can I use Memo offline?

Yes. Memo runs entirely on your device. The llama.cpp server provides local inference without internet. External providers (OpenAI, etc.) require internet, but the local model works fully offline.

Which models are supported?

Any GGUF-format model compatible with llama.cpp. This includes Llama 3, Mistral, Gemma, Phi, Qwen, and many others. You can browse and download models from the built-in Model Store, or use a model you already have.

What's the difference between Agent and Orchestra?

Agent mode uses a single LLM with tool access to autonomously complete tasks — like coding, file management, or research. Orchestra mode coordinates multiple agents, each with a specialized role, working together on complex multi-step projects. The conductor agent decomposes your request and assigns subtasks.

Can I use my own API keys?

Yes. Memo supports bring-your-own-key for 10 providers — OpenAI, Anthropic Claude, Google Gemini, xAI Grok, Groq, OpenRouter, Ollama, OpenCode Zen, OpenCode Go, and Kilo Code. Configure them in Settings → API Providers. Keys are encrypted at rest with AES-256-GCM.

Is Memo really free?

Yes. Memo is open source and free. There are no subscriptions, usage limits, or paywalled features. You provide your own API keys if you choose to use external providers.

How does memory work?

Memo generates embeddings for your conversations and stores them in a local vector database. When you ask a question, it retrieves the most semantically similar past memories and includes them as context for the LLM. This allows Memo to recall facts, preferences, and past discussions.

What happens when I'm offline?

Everything continues to work with the local model. Conversations, memory retrieval, agent tasks — all function without internet. External provider features are unavailable until you reconnect.

Can I sync between computers?

Yes. Enable Cloud Sync in Settings and set a strong passphrase. All data is encrypted before leaving your device. Multiple computers with the same passphrase will stay in sync automatically.

Can I use Memo with Claude Code or other Anthropic-compatible tools?

Yes. The Developer API Gateway (Sidebar → Developer) exposes a local, Anthropic-compatible endpoint — point ANTHROPIC_BASE_URL at it and Claude Code runs your own local model or your own provider API key behind the scenes, with full agentic tool calling for most providers.

What is Memo Swarm?

A beta feature (Sidebar → Swarm, after enabling Settings → Beta Features) that pools compute from several PCs to run one GGUF model too large for any single machine. One PC hosts the model and shares a room code; others join and lend compute without downloading the model themselves. It trades speed for capacity, and isn't available on macOS yet.

Can Memo do things on its own, without me asking?

Yes, in two ways. Routines let you schedule a prompt or agent task to run on a plain-language schedule. Proactive Learning notices habits and patterns and surfaces gentle nudges on its own — on by default, but fully disabled under Incognito and controllable (or fully off) under Minimal Mode and Settings → General.

Can I run Memo on a server, NAS, or Raspberry Pi instead of my desktop?

Yes, as of v3.5.5 — a dedicated server-only install (get-memo-server.sh), Docker/CasaOS images, a real 4-mode auth system, multi-account support, and a full memo config/remote/service/provider/agent/model CLI toolkit for managing it entirely over SSH, with no display required. See Self-Hosting and the CLI Reference.