Time tied to work
Desktop tracking against assigned projects and tasks, with notes, policy-defined screenshots and activity levels kept alongside the hours.
An AI-powered workforce-visibility platform for distributed teams that connects time, activity, screenshots, projects and approvals in one record — and writes up each person’s and team’s day every morning.
Agencies, software teams, BPOs and remote organisations usually keep time, activity, project and approval data in separate tools. Claresight connects them so managers can see the work behind logged hours, stop chasing updates, and catch scope creep or margin problems early.
Desktop tracking against assigned projects and tasks, with notes, policy-defined screenshots and activity levels kept alongside the hours.
Every morning Claresight writes up each person’s and team’s day from the tracked figures — and shows those figures next to it for checking.
Ask in plain words how a week looked or how a team’s day went; Clare only answers from data you already have access to.
Timesheet review and approval with work context, leave requests, schedules and late/early flags against real working days.
Admins, owners, managers, employees and clients each see the right slice of the workspace — nobody sees more than they should.
No keystroke recording, monitoring stops when the session stops, web addresses are never sent to AI, and employees can see their own data.
A modular NestJS API serves the web dashboard, a desktop agent for Windows and macOS, and Android and iOS apps, with live updates pushed to managers over WebSockets.
Tracking runs in a Tauri app that asks the operating system directly for idle time, the window in front and whether the screen is locked, so nothing is captured while it is. Offline time is kept in a local SQLite cache and synced when the connection returns.
Roles from platform admin to client are grouped in code into owners, managers and self-scoped people, and every query goes through one shared access service. Clients can have areas like screenshots or activity hidden from them.
The server works out each day’s figures first and the language model only words them. Summaries run as a background job after midnight in each organisation’s own time zone.
Clare answers by calling read-only tools that run under the asker’s own permissions, so it can never see more than they can. Usage is capped and metered against a monthly AI budget per organisation.
Stripe subscriptions with webhook replay and plan-based feature limits, and a retention job that deletes old screenshots and activity on each plan’s schedule.





