Search. Intelligently.
Lightning-fast search for macOS, powered by on-device machine learning. Find apps, files, messages, and more — all without compromising your privacy.
Lightning-fast search for macOS, powered by on-device machine learning. Find apps, files, messages, and more — all without compromising your privacy.
Open Vector and start typing. Apps appear instantly with strong keyboard shortcuts. No distractions, no shifting results — just consistent, predictable search.
Our custom machine learning model understands context. Type a person's name and see contacts, not dictionary definitions. Type an address and get maps, not web results.
Ask natural questions about your files and messages. "When was Joseph planning on visiting?" or "Which Swift file manages networking?" — Vector understands.
Never lose what you copied. Built-in clipboard history with instant search, all stored securely on your device. Access it anytime with a dedicated shortcut.
Custom indexing engine delivers near-instant file search results. Significantly faster and more reliable than Spotlight — no more waiting.
Leverages macOS 26's Liquid Glass effect with buttery-smooth animations. Customize the panel position and every data source to match your workflow.
Vector adapts to your day. The ambient mode intelligently surfaces what you need — calendar events in the morning, weather before you leave, maps when you're heading out. All based on context and time.
Search apps, check the weather, find addresses on maps, view your calendar, search contacts, and calculate with currency conversions — all from one powerful interface. Web results coming soon.
Ask natural questions like "When is Joseph visiting?" or "Which Swift file handles networking?" Vector uses on-device semantic search to understand meaning, not just keywords.
Your data stays yours. Vector processes everything on-device using local machine learning models. No cloud dependencies, no data collection, no compromises.
All embeddings and intelligence powered by local machine learning. Your searches never leave your Mac.
Semantic search and intelligent features powered by local ML models. No cloud API calls, no latency, no internet required.
Anonymous usage analytics via TelemetryDeck help improve Vector, but you can opt out anytime. No personal information is ever collected.
Adjust panel position, configure data sources, and tailor keyboard shortcuts to your preferences.
Vector is built for everyone. Choose what you'd like to pay — because great tools should be accessible to all.
View on GumroadmacOS 26+ • No subscription required
Vector bundles two ML models: the routing model, which determines which data sources to show based on a query, and the embedding model, which handles semantic intelligent search of iMessage and Files.
We do as much processing on the Neural Engine and aggressively optimize for memory. If you enable semantic search, then indexing could use up to 800MB of memory until it's done and then it would idle at less than 100MB usually. Every update has improvements to bring it down. The target is around 30MB idle or less. During indexing, the CPU might be used more than usual but this is temporary.
Quitting Vector and moving it to your trash should be enough. You'd probably want to delete this folder afterward since it can be a few hundred megabytes:
~/Library/Application Support/Vector
You just need an Apple Silicon Mac running at least macOS 26.0. I personally run on an M4 and M4 Max. You should have a decent amount of storage available for indexing. It might not perform as well on M1 and M2.
I really like fried rice, but I've still been trying to chase the high of the fried rice I had in Chinatown, NY. I also really like steak and milkshakes. I could live off shawarma.
If there is a major issue, it would usually be fixed within a few days. If there's a new feature that excites me, I'll add it within a week. I will make sure it's largely bug-free.
I mean, I guess? Not sure why you'd use Spotlight when this has more features and is more reliable. This isn't really meant to be a competitor to Raycast. This is much more for people that have been using Spotlight this whole time.