Jason Tyler Rodriguez

Independent project · not launched

DataDustr

Designing a more deliberate cleanup workflow

React / TypeScript / Next.js / Tauri / Rust

Try this demoFilter the sample inbox, then try Safe Mass Delete. Protected messages stay.
Data Dustr MVPInbox DustrWeb · Free
Inbox Dustrjason@example.com
✓ 7 emails loaded
Interactive sample
✧ SUGGESTED LABELS
+ Work+ Receipts+ Newsletters
Examples of labels suggested by the application.
All Emails viewSafe delete skips protected emails. This sample keeps 2 protected messages.
7 messages
Your annual subscription receiptStudio Billing
Protected
Your signed workshop agreementWorkshop Team
Protected
Last season’s design roundupDesign Dispatch
newsletter
Your weekly reading listProduct Weekly
newsletter
An older issue from your reading pileFrontend Notes
newsletter
A sale that ended last monthMarket Updates
promotion
Your expired weekend offerEveryday Goods
promotion

Review the sample messages, then try Safe Mass Delete.

Adapted from the live Inbox Dustr interface.Reduced interactive demo · synthetic messages · local changes only

Independent product initiative

From digital clutter to a reviewable decision.

An independent file and inbox organization product, built around reviewing suggestions before changing data. The business has not launched.

My role
Independent product design and engineering
Scope
File Dustr, Inbox Dustr, and a supporting service website
Runtime
Web demonstration; desktop integration through Tauri

Inside Inbox Dustr

These screenshots were captured from the live web application on October 18, 2026 using Gmail mock data. Overview summarizes the inbox; Insights surfaces patterns; Organize groups and protects messages; Unsubscribe collects recurring senders.

Try the live web demo ↗

Overview: inbox totals, categories, and cleanup opportunities
Overview: inbox totals, categories, and cleanup opportunities
Insights: engagement patterns and oldest clutter
Insights: engagement patterns and oldest clutter
Organize: suggested labels, filters, and safe deletion
Organize: suggested labels, filters, and safe deletion
Unsubscribe: recurring senders and mailing-list actions
Unsubscribe: recurring senders and mailing-list actions

Watch the four-screen GIF tour ↗

Live application captures · mock data. The interactive preview above adapts a smaller sample for this portfolio.

The interaction comes before the operation

The review queue groups suggestions and explains rule matches, keeping the decision separate from the operation that changes data.

  1. ScanLoad files or demonstration data
  2. SuggestClassify and explain a rule match
  3. ReviewLet the person inspect an action
  4. ExecutePass the request to a runtime adapter
Implementation map, not an application screenshot. Based on the file workflow and runtime boundaries.

Decision 01

Keep the domain separate from the runtime

Shared file types and workflow helpers feed the React interface. A platform adapter separates web demonstration behavior from native filesystem and email operations.

The web demo can run without filesystem access; the desktop boundary needs its own integration tests.

Decision 02

Make policy explicit

One policy module controls edition and feature availability. Unit tests cover policy, cleanup heuristics, and workflow state.

The current entitlement is local state. It is not verified payment or a production licensing service.

What still needs work

End-to-end behavior needs hardening: size and extension do not prove duplicate content; success must follow execution results; undo must restore real data before promising recovery.

The web edition uses demonstration data. Native cleanup, account access, and cross-platform releases are not presented as production-ready.

A single product story

The service website supports this same initiative. Business validation, commercial licensing, and broader integrations remain future work.

DataDustr service site with its dark background, red accents, service proposition, and consultation call to action.
Source-based service-site illustration

A separate service site explains the offer and provides contact and consultation paths. It adapts an existing starter; the business has not launched.

Application source is private. The illustration uses invented sample data; no account data is included.