Private Semantic Search
for macOS
Search your files by meaning, not just keywords.
100% local. Zero tracking. Your data never leaves your Mac.
Our quarterly goals for Q1 focus on expanding the team and launching the new product line. Key discussion points include...
During the board discussion we reviewed our annual targets and agreed on the quarterly milestones for each department...
Why Another Search Tool?
Built-in search hasn't evolved. Third-party tools compromise your privacy.
Spotlight is Keyword-Only
Looking for that document about 'quarterly goals'? Better remember the exact word 'quarterly' — because Spotlight won't find it if you search 'Q1 objectives'.
Cloud Search Exposes Your Data
Every search query sent to the cloud is metadata about your work, habits, and thoughts. Privacy policies change. Servers get breached.
AI Search Doesn't Work Offline
Internet down? Too bad. Most AI search tools require constant connectivity and monthly subscriptions.
Semantic Search That Never Leaves Your Mac
Mür combines cutting-edge AI with absolute privacy. Your thoughts stay on your device.
Semantic Understanding
Search by concept, not exact keywords. Powered by Rime vector search + local embeddings.
"Why?" Explanations
See exactly why each result matched — relevance score, matched concepts, preview text.
100% Local
All processing on your Mac. No network calls. No tracking. Works completely offline.
Lightning Fast
Sub-100ms search latency. Native Metal GPU acceleration for Apple Silicon.
Auto-Watch Folders
Monitors your Documents, Desktop, Downloads for changes. Auto-reindexes in the background.
Open Source Kernel
Built on Rime — embeddable vector search (Rust crate). Use it in your own projects.
Explore Rime →Local vs Cloud
See exactly how Mür keeps your data private while delivering powerful AI search.
Cloud Search
Mür
PrivatePowered by Rime (vector search) + NCPPCN (local inference)
Embed Rime in Your App
Mür is built on Rime — an embeddable vector search kernel. Think "SQLite for semantic search." Add Rime to your Rust project with one line.
Explore Rime on GitHub[dependencies]
rime = "0.1"
[features]
default = ["simd"]
simd = ["rime/simd"]use rime::Index;
let index = Index::new()
.embedder("local")
.build()?;
let results = index
.search("semantic query")
.limit(10)
.execute()?;
for result in results {
println!("{}: {}",
result.score,
result.text
);
}Ready in 3 Steps
Download, index, and search. Your private semantic search starts in minutes.
Download
Grab the .dmg for macOS 13+. ~50MB.
Index
Choose folders to search. Auto-detects Documents, Desktop, Downloads.
Search
Press ⌘ ⌥ Space and type what you need.
Requirements: macOS 13+ (Ventura) · Apple Silicon or Intel · ~500MB for models
Free and open source. No account required.