Voyager
Find files by context, not just folders.
An invite-only beta for Voyager is now live
We are starting the first invite-only beta on January 26, 2026. We will keep it invite-only until we announce otherwise.
In this phase, we are validating one core idea: making macOS Smart Folders easier to create and use, without relying on a complex settings UI.
Getting access
This beta is invite-only. If you are selected, we will send you everything you need to install and activate Voyager by email.
That email includes:
a download link
a beta code tied to your email address
If you have not received an invite yet, we will invite users sequentially based on the waitlist signup order. We may adjust the order depending on product validation needs and operational constraints.
Collections and filters
Collections are saved filters that show matching files in one place, without moving anything. You can still browse folders as usual, then use Collections as a rule-based view of those files.
In this beta, we focus on the full Collections flow from start to finish:
Compose a filter in plain language
Review the filter Voyager suggests
Add, remove, or tweak conditions until it matches what you want
Save the collection and reuse it anytime
Example prompts
Find PDFs modified in the last 30 days.
Show me spreadsheets from 2024 in Reports.
Files name contains “report”
What’s next
This beta is focused on Collections. Next, we’re exploring:
Chat with your files and request actions
Modernized browsing UX
File management automation setup
These priorities are not fixed. The order can change based on beta feedback, real usage patterns, and participant preferences.
Privacy
Privacy is a core product principle.
During the beta, we may need limited diagnostic information to improve stability and quality. We will be transparent about what is collected, when it is collected, and why it is used.
Communication channels
Email: support@voyager.fm
Discord: https://discord.gg/z3cFMMGTuX
Even if you have not been invited yet, you are welcome to join the channel and share feedback.
5 Comments
Finding a specific file often feels like a chore when you have to remember exactly which folder you buried it in months ago. It is a massive relief to just describe what you need in plain language instead of clicking through endless subdirectories and complex search settings. Since you are focusing on making macOS Smart Folders more intuitive how do you handle privacy when processing those plain language prompts locally on the machine?
Great question.
Privacy is one of the main things we’re being careful about.
The app is designed as a local-first macOS app. File metadata and collection logic stay on the user’s machine, and the goal is to avoid uploading personal file contents.
For the beta, we’re being especially conservative here because file systems contain a lot of private context.
Long term, I want Voyager to feel more like a local workflow layer than a cloud search product.
Keeping the file metadata and logic local is a major selling point because people are naturally protective of their file systems and don't want a "cloud search" product indexing their private documents.
Focusing on the local workflow layer gives you a distinct advantage over generic AI tools because you can offer the speed and security of a native macOS experience without the latency or privacy risks of an external server.
I emphasize this kind of "privacy-first" authority in my work with high-tier PR and media placement where we highlight a brand's technical integrity to secure features on major tech outlets.
Are you considering adding support for specific tags or file colors in the plain language prompts so users can filter by their existing organization systems?
Yes, that is very much on our radar. One of the advantages of keeping search local is that we can make plain language search work with the organizational systems people already use on their Mac, without sending private file metadata to the cloud.
Today, we are better positioned to support existing tag names in natural language prompts, since those map cleanly to local metadata and fit our privacy first model well. File or tag colors are a bit more nuanced, so that is something we would likely treat as a next step rather than overpromise immediately.
Our goal is to make search feel natural while still respecting the structure users have already built into their files, especially in a way that stays fast, native, and privacy preserving on macOS.
The distinction between supporting existing tag names now versus overpromising on color filters is exactly the kind of honest product scoping that builds long-term user trust — beta users notice when a team doesn't overpromise.
The privacy-first local approach is actually a stronger differentiator than it might seem right now. As more AI tools get scrutinized for data collection, a native macOS tool that never touches a server becomes a genuinely compelling story — especially for power users like developers, lawyers, and designers who are most protective of their file systems.
Are you planning to highlight that local-first architecture as a core part of your positioning when you move beyond invite-only?