SaaS Development • Native Desktop Agent • In-House Infrastructure
TimoDesk - Time Tracking Software For Global Teams
Overview
TimoDesk is time tracking software for distributed teams. A native desktop agent and browser extensions record working hours, periodic screenshots and activity levels; a web dashboard turns that stream into worklogs, timesheets and productivity reports managers can act on. THESOFTKING delivered the product end to end - consultancy, product planning, UI/UX, the web application, the tracking clients, the storage and server infrastructure we built and operate in-house, and ongoing support. The defining constraint was price: every feature ships on a single plan at one dollar per user per month, which meant the architecture had to make screenshot storage, normally the most expensive part of a monitoring product, cheap enough to disappear into that number.
Industry
- SaaS
- Workforce Management
- Remote Work
- Productivity
Service we provided
Technology Used
A Complete Time Tracking Platform for Global Teams
TimoDesk is time tracking software built for teams that no longer share an office. A lightweight desktop agent runs on each team member machine and records working hours, periodic screenshots and activity levels. Browser extensions cover the same ground inside the browser, so people working in web tools all day can start and stop a timer without switching windows. A web dashboard turns that raw stream into worklogs, timesheets, project breakdowns and productivity reports a manager can act on.
THESOFTKING delivered the entire product: consultancy and product planning, UI/UX design, the web application, the native desktop agent, the browser extensions, the storage and server infrastructure underneath it all, and the ongoing release and support cycle that keeps it running today.
Automatic time tracking from a native desktop agent that keeps recording when the connection drops and syncs the moment it returns.
Screenshot monitoring at configurable intervals, stored and served entirely from storage infrastructure we built and operate ourselves.
Activity analysis that converts idle time, input rates and application use into a productivity score per person, project and day.
Time calendars, worklogs and exportable reports that compress a month of tracked activity into a decision in a few clicks.
Browser extensions that start and stop timers inside the web tools teams already work in, with no context switch to a separate app.
One plan at one dollar per user per month, every feature included, made possible by infrastructure designed around the cost of storage.
The Challenge: Screenshot Monitoring at One Dollar Per User
The price came first. TimoDesk had to ship every feature on a single plan at one dollar per user per month, with no tiers and nothing held back for an enterprise upsell. That is a fraction of what the established names in employee monitoring software charge, and it meant the entire architecture had to be designed backwards from that number rather than the other way around.
Screenshot monitoring is the part of the product that makes this hard. A fifty person team on a ten minute capture interval generates well over fifty thousand images a month, and every one of them has to be stored, retained, thumbnailed and served back into the dashboard on demand. Rented through a third party object storage service, billed per gigabyte stored, per gigabyte transferred and per request, that single feature would have consumed the whole subscription before a line of the product was paid for. Every competitor solves this by charging more. We had to solve it by charging less.
The tracking clients were the second problem. Monitoring software has a deserved reputation for being heavy, and an agent that drains battery or spins up fans undermines the product it is meant to support: the people being measured are the same people who can quit the app. It had to run all day, capture continuously, survive an unreliable connection, and stay quiet enough that nobody notices it is there.
The Solution: Own the Storage Layer, Not Rent It
Rented object storage is priced for a general workload: unpredictable reads, global access patterns, durability guarantees far beyond what a screenshot from last Tuesday needs. The TimoDesk workload is none of those things. It is narrow and completely predictable. An image is written once by an agent, read rarely and by one person, kept for a fixed window, then dropped. Paying general purpose prices for a special purpose workload is exactly where the money in this category disappears.
So we did not rent it. TimoDesk writes to storage infrastructure we designed, provisioned and operate ourselves, sized specifically for the shape of that workload. It is the single decision the rest of the product economics rests on, and it is what makes one dollar per user per month a sustainable price rather than a promotional one.
Infrastructure: Built and Operated In-House
Infrastructure is not a supporting detail on this project. It is the product decision that every other decision hangs off, and THESOFTKING owns all of it: architecture, provisioning, deployment, monitoring, capacity planning and the ongoing operations that keep it running.
The ingest path is where most of the saving is made. Captures are compressed and thumbnailed the moment they arrive rather than on the way out, so the dashboard browses cheap derivatives and only pulls a full resolution image when someone explicitly asks for one. Because the overwhelming majority of stored screenshots are never opened by anyone, that one change removes most of the bandwidth a naive implementation would burn. Retention is enforced automatically on a schedule, which keeps total volume flat instead of compounding month over month as the customer base grows.
The commercial effect is a change in the shape of the cost, not just its size. Rented storage is a variable bill that scales with every customer won, which means growth and margin pull against each other. Owning the tier turns it into planned capacity we control, so adding a customer adds load to hardware already paid for rather than a proportional invoice from a third party. That is what lets the price stay flat as the platform grows.
There is a second reason to own it. TimoDesk holds screenshots of people at work, which is about as sensitive as operational data gets. Knowing exactly which machines that data sits on, who can reach them, and when it is deleted is a security posture rather than a line item, and it is not something you can fully delegate to a storage vendor.
The Tracking Clients: From Electron to Native, Plus the Browser
The first release shipped in Electron. With four developers and a two month window to get a working product in front of real teams, it was the right call: one codebase, fast iteration, and a shell the team already knew. It got TimoDesk into the market and, more importantly, it produced real telemetry about how the agent behaves on machines we do not control.
That telemetry is also what retired it. An Electron shell carries a browser runtime it barely uses, and on the low specification laptops common across the teams TimoDesk serves, the memory footprint was doing real damage to the experience. So we rewrote the agent natively. The native build starts faster, holds a fraction of the memory, captures without the frame hitches users had learned to expect from monitoring tools, and talks to the operating system directly for idle detection and screen capture instead of going through an abstraction. Shipping in Electron first and rewriting once we had evidence was not a detour: it bought two months of market feedback that told us precisely which parts of the agent were worth optimising.
Browser extensions complete the picture. A large share of the work TimoDesk measures happens entirely inside a browser tab, and asking someone to alt-tab to a separate window to start a timer is exactly the kind of friction that stops a tracker being used honestly. The extensions put start and stop where the work already is. Both surfaces share the same offline behaviour: they queue locally when the network is unavailable and reconcile when it returns, so a bad connection costs a user nothing and a manager sees no gaps.
Visual Design: Making Monitoring Feel Fair
A product that photographs people at work has to be careful about tone. Designed badly, this category feels like surveillance, and the interface alone can make identical data read either as evidence against someone or as a straightforward account of a working day. TimoDesk is deliberately matter of fact. The palette is light and neutral with a single accent reserved for actions and live state, charts are plain and legible at a glance rather than decorative, and screenshots are presented as a calm timeline rather than a wall of captures.
The dashboard is dense with data but structured so the common questions, who is working now, where did this week go, is this project on budget, are answered on the first screen without a filter or a click. The tracking clients take the opposite approach: one window, one visible control, and nothing to configure day to day.
The same thinking shaped billing. TimoDesk runs on Stripe but deliberately does not store cards or charge them automatically. Invoices are generated and customers choose to pay them. For a product whose whole job is to record what people do at their desks, not holding a card on file is a small decision that says something about how the rest of the data is handled.
Conclusion: Price as an Engineering Outcome
Four developers took TimoDesk from a brief to a first release in two months, and the platform has been in continuous development and support since. It now serves more than five hundred teams, has tracked over a hundred thousand hours of work, and teams using it report an average thirty eight percent improvement in measured productivity.
The interesting part of this project was never the feature list. Time tracking is a well understood category with capable incumbents. What TimoDesk proves is that the price a product can charge is an engineering outcome, not a marketing one. By building and operating the storage tier instead of renting it, and by rewriting the tracking agent once the market showed us where it hurt, the most expensive feature in the category stopped being a constraint on what the product could cost. That is the difference between a cheap product and a product that can afford to be cheap.