FAQ
Don’t see your question? Ask in the Q&A.
Are any queries sent to the cloud?
Section titled “Are any queries sent to the cloud?”Only if you choose a cloud AI provider and the notes involved are not marked private. Search and the graph run entirely locally. With a local provider such as Ollama, everything stays on your machine.
Even with a cloud provider configured, you can mark notes or folders private. Those are blocked from being read, embedded, or sent to any untrusted provider. See the privacy model.
What are the limitations?
Section titled “What are the limitations?”Worth knowing before you start:
- Quality depends on the model. Models vary in reasoning, tool use, and embedding quality, so results differ between them.
- Quality depends on your vault. Answers improve when notes are well-structured and each covers a coherent topic. Notes mixing many unrelated subjects retrieve poorly.
- The assistant can be wrong. If relevant notes don’t exist, or the model misreads them, you’ll get an unsatisfying answer. Rephrasing the query or describing the context in more detail usually helps.
What models do you recommend?
Section titled “What models do you recommend?”Use the most capable model your provider offers. Agent quality depends heavily
on reasoning and tool-use ability. For a fully local setup, pair a capable
Ollama chat model with a strong embedding model such as qwen3-embedding:0.6b,
which is multilingual and runs comfortably on a laptop.
For current rankings — and which benchmarks to weight when reading them — see Choosing a model.
Does it work with non-English vaults?
Section titled “Does it work with non-English vaults?”Yes. Response quality varies with the model and the language it reasons in, and strong multilingual embedding models give the best retrieval results. More UI translations are on the way.
What’s next?
Section titled “What’s next?”What’s planned, in progress and done is tracked on the public project board.