IONICWEBCREATOR
Case Studies
LiveDreamDesk

DreamDesk

Turning a scattered set of productivity tools into one calm, multi-tenant workspace that stays fast as it fills up.

  • Productivity
  • 2025
  • 6 min read
Overview

Project overview

DreamDesk wanted the focus of a single tool with the breadth of a suite — tasks, notes, and scheduling on one surface. We designed a multi-tenant workspace platform that stays quick and keyboard-driven no matter how much a team pours into it.

Client
DreamDesk
Industry
Productivity
Category
SaaS Platforms
Role
Product design & engineering
Year
2025
The challenge

The business challenge

Teams were stitching together three or four apps and losing context in the gaps. A unified workspace only helps if it stays fast under real data — the moment it feels heavy, people go back to their tabs.

Discovery

Discovery & strategy

Where we started before writing code — framing the problem, the constraints, and the plan.

  1. Mapped the daily loop

    We shadowed how teams move between tasks, notes, and calendar in a day, and designed the surface around that loop rather than a feature list.

  2. Chose tenancy early

    Multi-tenancy touches every query and every permission, so we settled the isolation model before building features on top of it.

  3. Committed to keyboard-first

    Power users live on the keyboard, so command palette and shortcuts were requirements from day one, not a later polish pass.

Architecture

System architecture

A typed, end-to-end platform: a Next.js client talking to a tRPC API over a Prisma-modelled Postgres, with Redis caching hot workspace data so switching contexts feels instant.

  • Client

    A keyboard-driven workspace UI with optimistic updates and a global command palette.

    • Next.js
  • API

    End-to-end typed procedures over tRPC, so the client and server never drift.

    • tRPC
  • Data

    Row-level multi-tenant Postgres modelled with Prisma, hot paths cached in Redis.

    • Prisma
    • PostgreSQL
    • Redis
Decisions

Design decisions

The trade-offs worth explaining — what we chose, and why it was the right call for this build.

  • tRPC for a single type boundary

    One type flows from the database to the button, so a schema change surfaces at compile time instead of as a runtime bug in a client that got out of sync.

  • Row-level tenancy over separate databases

    Shared schema with strict row-level scoping kept operations simple while still isolating every workspace's data.

  • Optimistic UI with server reconciliation

    Actions apply instantly and reconcile with the server, so the workspace feels immediate without lying about failures.

Stack

Technology stack

Every layer of the system, and the tools chosen to build it.

Frontend
  • Next.js
Backend
  • tRPC
Data
  • Prisma
  • PostgreSQL
  • Redis
Features

Key features

The parts of the build that carried the most product weight.

  • Unified surface

    Tasks, notes, and scheduling in one navigable workspace.

  • Command palette

    Everything reachable from the keyboard in a keystroke.

  • Real-time sync

    Changes propagate across a team's sessions without a refresh.

  • Granular permissions

    Per-workspace roles that scope every read and write.

Results

The outcome

DreamDesk shipped a workspace that stays fast as teams fill it, replacing several tools with one surface people actually keep open all day.

Tools replaced
4→1
Retrospective

Lessons learned

What the engagement taught us, kept honest rather than tidy.

  1. Settle tenancy first

    Deciding isolation early saved us from reworking every query later.

  2. Fast beats featureful

    People adopted the surface because it stayed quick, not because it did more.

  3. Types are a feature

    The end-to-end type boundary caught whole classes of bugs before release.

Related

Other work with a similar shape — by industry, stack, or the problem it solves.

  • HOHouse of Velore
    Live

    House of Velore

    Fashion & Retail

    A luxury fashion storefront built for a considered browse and a frictionless checkout — fast on every device and structured so the catalogue scales with the collection.

    • Next.js
    • React
    • TypeScript
    • PostgreSQL
    • AWS
  • FSFuture Screen
    In Development

    Future Screen

    Digital Media

    A digital-signage control plane for scheduling and pushing content to screens in the field — built for reliability and remote management at scale.

    • Next.js
    • Node.js
    • PostgreSQL
    • Docker
  • DDeco4U
    Live

    Deco4U

    Home & Décor

    A home-décor commerce experience with rich product discovery and a checkout tuned for conversion — headless, fast, and easy for the team to merchandise.

    • Next.js
    • TypeScript
    • PostgreSQL
    • Tailwind CSS

Have a problem worth solving well?

If the thinking above matches the standard you're after, tell us what you're building. We'll be honest about fit and outline a concrete first step.