A 100% free, private operating system for your household or team — tasks, projects, calendar, notebooks, routines and a built-in AI assistant, all running on a machine you already own. Nothing ever leaves your LAN.
No sign-up, no SaaS dashboard, no credit card. npx paperr fetches the app, then install:all creates server/.env and auto-generates your JWT secrets — no manual copy step. paperr binds to 0.0.0.0, so every phone, tablet and laptop on your network reaches the same instance.
npx paperr
Needs Node.js 22.5+ for the built-in node:sqlite module. The bundled AI server auto-installs if Python 3 is present — otherwise skipped safely, or point dotAi at Ollama / LM Studio.
$ npx paperr # fetches paperr into ./paperr $ cd paperr $ npm run install:all # deps + auto-writes .env $ npm run dev # API :3000 · client :5173
$ git clone https://github.com/biswasprateek/paperr.git $ cd paperr $ npm run install:all # deps + auto-writes .env $ npm run dev # API :3000 · client :5173
$ npm run build # bundles the React app $ npm start # Express serves API + app :3000 # then open from any device on the LAN $ open http://<your-ip>:3000
# builds the client, then serves everything on :3000 > ./start-paperr.ps1 # Windows $ ./start-paperr.sh # macOS / Linux > ./start-paperr.ps1 -NoBuild # skip rebuild
Cloud apps stop working when the internet drops and lock your life behind a subscription. paperr runs on hardware you control — no accounts, no telemetry, no cloud.
No subscription, no per-seat pricing, no "pro" tier. Clone it, run it, done. Apache 2.0, yours to keep and adapt.
Runs on a machine you already own; data lives in SQLite files on your network. No accounts you don't control, no telemetry, no cloud.
A bundled LiteRT Server AI model runs on CPU in under 2GB of RAM — and puts your iGPU or GPU to work when there's one, scaling to whatever hardware you have. No cloud; auto-offloads when idle. Or bring Ollama / LM Studio.
Create Spaces — a Family space for home, a Team space for work — each with its own members, Areas and fully isolated data.
Set paperr up once on a single machine, and the whole family or team signs in from their own phone, laptop or tablet — no per-device install.
Changes sync instantly across every device over Socket.io. Check a task on your phone, watch it tick on the wall dashboard.
Create, assign, schedule and complete tasks with sub-tasks, tags, comments, attachments and a markdown scratchpad. Due dates plus explicit start/end time blocks that land on the calendar, recurrence by weekday, and smart Today / Overdue views with drag-and-drop and bulk actions.
A customizable, swipable widget board — add the widgets that matter and arrange your home screen. A separate shared Hub gives the whole space a collaborative landing page with Sticky Pads, a per-member calendar and everyone's routines side-by-side.
Pick a task, go fullscreen, and get nothing else until you're done. A distraction-free Pomodoro or count-up timer runs beside the task detail, with a "still working?" check-in every 15 minutes and time logged per task — plus a shared widget showing who's heads-down right now.
Day / Week / Month / Agenda views, events with attendees, per-day habit rings, weather and an "up next" rail.
Time-anchored habits that auto-reset daily — Protocols, a chronological Day Arc, and streak-tracking Progress.
Full project management for the bigger stuff — Board, List and Phases views, per-phase milestones, a Watch toggle for notifications, per-project members and a filtered activity feed.
An Obsidian-style knowledge base of markdown notes — recipes, manuals, house docs. A shared brain for your space.
Pomodoro, Stopwatch, Timer, guided Breathe, Meditate, Ambient Sound, a flip-clock with alarms, and Good Thoughts — plus an Analytics page for streaks, trends and a Pomodoro heatmap.
Turn any screen into a digital photo frame — your own photos or a curated public-domain art set, synced server-side and played fullscreen.
dotAi runs on paperr's bundled AI server — a small, efficient AI model that installs itself and runs on as little as CPU in under 2GB of RAM, uses your iGPU or GPU when available (scaling to the hardware you have), then auto-offloads when idle. Prefer your own model? Point it at a local Ollama or LM Studio and your conversations and data still never leave your network.* And it uses tool-calling to act on your data, not just chat. Don't love a reply? Regenerate it, then flip between every alternate response with the arrow buttons underneath.
Background agents run on a schedule and surface calm, dismissable insight cards on the Agent Hub — never popups, never applied automatically.
| Agent | What it does |
|---|---|
| Morning Brief | A short daily summary of today's tasks, overdue items and events. |
| Reschedule Advisor | Spots overdue tasks and proposes smart new dates — one tap to accept. |
| Priority Focus | Surfaces what matters most on a stacked day. |
| Workload Spread | Flags overloaded upcoming days and proposes redistributing tasks. |
| Bulletin Board | A space-wide digest of completed tasks and habits. |
Or build your own Custom Agents — a name, a schedule, and free-text instructions — testable live before you save.
* Aiming it at a cloud endpoint — OpenAI, OpenRouter and the like — is opt-in, and does send that data to a third party.
paperr isn't designed only for neurodivergent users, but a lot of what makes it calm makes it genuinely easier to live with if you have ADHD, are autistic, or just run low on executive fuel — supporting everyone in your family or team, side by side.
Deep Work Mode goes fullscreen on a single task with a gentle "still working?" nudge. No tab, no feed, no next thing.
Today / Overdue views and the Morning Brief surface what matters, so a stacked day doesn't start with a blank-page decision.
The Reschedule Advisor proposes new dates for overdue tasks with one tap. Nothing auto-applies — every suggestion waits for your yes.
Dismissable insight cards instead of popups. Calm surfaces, quiet defaults, nothing clamouring for attention.
Routines auto-reset daily and lay the day out on a Day Arc, plus reusable named timers ("Laundry", "Leave in 10") for time-blindness.
Guided Breathe, Meditate and Ambient Sound with an optional before/after mood check, so you can see a session helped.
A master Reduce Motion switch (on by default if your OS already asks for it) plus per-effect toggles for celebrations, weather icons, widget jiggle and alarm shake — kill just the one thing that bothers you.
No separate apps to install. The mode is auto-detected from screen size and touch support on first load, remembered after, and switchable anytime.
Sidebar navigation, dense multi-column layouts and a slide-out dotAi chat drawer for the full workstation experience.
A bottom-nav, touch-first shell built around the customizable widget Home board — your whole home in a pocket.
The full app in a touch-optimized nav-rail shell — a wall-mountable dashboard for the kitchen or office, sized for a mounted display.
Dev behind paperr · biswasprateek
paperr is a solo project, built by a dad who got tired of watching tasks, routines,schedules and to-dos scatter across different apps for his family. I'm obsessed with building one system that actually captures and tracks everything happening for busy families and teams — without a stack of cloud subscriptions or handing over your data to run it. This is also where I get to push what Small Language Models (SLMs) can do on hardware people already own, bringing real everyday value out of local AI. Built with love in New York.
Clone the repo, run two commands, and open paperr from any device on your network. Your home, organized — and your data stays yours.