paperr is a private, self-hosted operating system for your household or team — tasks, projects, calendar, notes, routines, focus tools, a shared wall dashboard, and a built-in AI assistant, all running on your own network. Nothing ever leaves your LAN.
100% free. Your data never leaves your network. A tiny but powerful AI that runs on your machine.
Quick Start · Features · dotAi · Modes · Tech Stack · Contributing
A calendar app answers one question: what's on today? Real life is bigger than that. You've got tasks to assign, projects to run, groceries to remember, house docs to keep, habits to hold, and a family or team to keep in sync — and most tools make you rent a different cloud subscription for each one, hand over your data to do it, and lose it all the moment the internet drops.
paperr replaces the whole stack with one app you host yourself — and it costs nothing.
- 🆓 Free — no subscription, no AI credits caps, no "pro" tier. Clone it, run it, done.
- 🔒 Truly private — runs on a machine you own; data lives in SQLite files on your machine. No accounts you don't control, no telemetry, no cloud. It never leaves your walls.
- 🤖 Local, private AI — dotAi runs on paperr's bundled AI server (based on LiteRT): a small, efficient Gemma E2B model that installs itself on
npm installand runs on CPU alone, under 2GB of RAM — the laptop or Mac you already have, no GPU or special hardware needed. It auto-offloads from memory when idle, so it costs you nothing while you're not using it. Or bring your own Ollama / LM Studio. Your tasks and projects are read and edited by a model on your hardware — nothing is sent to anyone else's servers unless you deliberately point it at a cloud provider. - 🗓️ Not just a digital calendar — tasks, projects, lists, notebooks, routines, focus & wellness apps, a photo Frame, and a shared Hub, all in one place and all talking to each other.
- 🏠 Built for a household or a team — create Spaces (Family or Team) with their own members, areas, and data.
- 🌐 One install for the whole family or team — set paperr up once on a single machine on your network, and everyone signs in from their own phone, laptop, or tablet — no per-device install, no per-person subscription.
- 📺 Turn a spare tablet into a shared screen — an old iPad, Fire tablet, or Android slate becomes a wall-mounted, always-on dashboard for the whole family or team — Home board widgets, Frame photos, and Hub, always visible.
- 📱 Every screen, one app — a polished desktop layout, a touch-first phone experience, and a wall-mountable tablet dashboard.
- ⚡ Real-time everything — changes sync instantly across every device via Socket.io.
- 🧠 Built to reduce friction for every kind of brain — calm, one-thing-at-a-time defaults that genuinely help if you have ADHD, are autistic, or just run low on executive fuel. More below.
| Home board |
|---|
 |
| Hub |
|---|
 |
| Tasks |
|---|
 |
| Lists |
|---|
 |
| Project View |
|---|
 |
| Month | Week |
|---|---|
 |
 |
| Routines |
|---|
 |
| Notebooks |
|---|
 |
| Frame |
|---|
 |
| Agents |
|---|
 |
| Apps |
|---|
 |
| Analytics |
|---|
 |
| Settings |
|---|
 |
| Home board — Dark Mode |
|---|
 |
Organize life into separate Spaces — a Family space for home, a Team space for work. Each space gets its own members, predefined Areas (Kitchen, Garage, Office… or Development, Design, Marketing…), and fully isolated, private-by-default data.
More on Spaces & accounts
Self-service registration (with optional quick-login PIN and avatar) and a Browse Spaces screen let people discover and request to join a space, with admin approval. Notebooks, Projects, Lists, and Custom Agents are private to their creator by default, with a one-tap toggle to share them with the whole space. Space admins can permanently delete a space (type-to-confirm) from the Edit Space screen. New spaces end with an optional "Set up your apps" step to enable the bundled AI server and add curated Frame photo and Good Thoughts collections. Every new space is seeded with type-appropriate starter lists, a starter routine, and a "Welcome to Paperr" onboarding project that tours the app.
The heart of paperr. Create, assign, schedule, and complete tasks with:
- Sub-tasks, tags, comments, and file attachments
- A markdown notes scratchpad per task
- Due dates and explicit start/end time blocks that show up on the Calendar
- Recurrence, including specific weekdays
- Today and Overdue smart views, assignment across household/team members, drag-and-drop organization and bulk actions
- A little confetti/emoji/toast celebration when a task or habit gets completed — tune individual effects down or off in Settings → Motion
Pick a task, go fullscreen, and get nothing else until you're done. A distraction-free Pomodoro/count-up timer runs alongside the full task detail and its sub-tasks, with a "still working?" check-in every 15 minutes and session time logged per task — plus a shared widget showing who on your team is heads-down right now.
Lightweight checklist-style lists with custom icons and an emoji picker — groceries, packing lists, honey-do lists. Items can optionally link to full tasks or projects.
Full project management for the bigger stuff — home renovations, trips, work initiatives:
- Board, List, and Phases views (a step tracker with per-phase progress)
- Phases can hold named Milestones, both manually markable complete
- A Watch toggle so members get notified when a watched project's tasks complete or get blocked
- Per-project members and a filtered activity feed
A real calendar with Day / Week / Month / Agenda views, events with attendees, habit time-of-day dividers and per-day habit rings alongside tasks, a mini month picker, quick-add box, per-day weather, and an "up next" rail.
An Obsidian-style knowledge base — organize Notebooks of full markdown notes. Recipes, manuals, meeting notes, house docs — a shared brain for your space.
Time-anchored habits that auto-reset daily, inspired by the "biological arc" idea:
- Protocols — goal buckets (Morning Optimization, Sleep, Fitness…)
- Day Arc — a chronological timeline of the day's habits
- Progress — streaks and per-protocol stats
A built-in suite of focus tools on the Apps page — Pomodoro, Stopwatch, Timer, guided Breathe, Meditate, Ambient Sound, a flip-clock Clock with alarms, and Good Thoughts (cycling affirmations/quotes/mantras you author yourself). Run them full-screen or pin any of them to your Home board.
Build your own collections of affirmations, quotes, or mantras — custom color, icon, category, and any text you write yourself, not just presets. Each collection can be time-bound (morning, afternoon, evening…) so the right thought shows up at the right part of the day.
More on Good Thoughts
Cycle them in a dashboard widget, or interleave them as text cards into Frame's fullscreen photo slideshow. New spaces can also add a curated starter collection (e.g. Stoic Quotes) in one click from the Apps setup step instead of starting empty.
Turn any screen into a digital photo frame. Add the photos you actually want — family pictures, vacation shots, whatever you upload — auto-downscaled and stored server-side so they sync to every device on your network, then play fullscreen with configurable transitions.
More on Frame
Prefer not to upload your own? Start from a hand-curated collection of public domain art imported in the background instead — either way, a calm, ad-free way to fill an old tablet or spare monitor between focus sessions.
A customizable, swipable widget board — add the widgets that matter (tasks, calendar, weather, projects, routines, activity, focus timers, Good Thoughts, Time Progress) and arrange your perfect home screen. The AI Agent widget surfaces live insights from any agent of your choice right on the board, so dotAi's proactive suggestions are one glance away instead of a page visit. A separate shared Hub board gives the whole space a collaborative landing page with Sticky Pads (corkboard notes), a per-member calendar, a shared list, and everyone's routines side-by-side.
A personal notification feed (project comments, task completions/blocks on things you watch) in the header bell, alongside a real-time activity feed of everything happening across the space.
Trigger a manual database snapshot or schedule automatic backups (daily/weekly/monthly with configurable retention) from Settings, with one-click restore (auto-snapshotting first) and delete.
Chat with dotAi anytime from the slide-out drawer to create, edit, and query tasks, projects, events, and more. Don't love a reply? Hit regenerate for another take, and flip between every alternate response with the arrow buttons underneath it.
Beyond chat, paperr runs background agents — on a schedule, via node-cron — that surface helpful, dismissable insight cards on a dedicated Agent Hub page:
| 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 daily/weekly digest of completed tasks and habits |
Cards are calm by design — no popups or spam. Dismiss, snooze, or approve with a tap. Anything an agent proposes (rescheduling, completing, creating a task) is queued for approval, never applied automatically. Every pre-built agent also has a Run Now button on the Agent Hub page to trigger it on demand instead of waiting for its schedule.
You can also build your own Custom Agents — a name, a schedule, and free-text instructions — which run the same tool-calling loop and can be tested live before saving.
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 — so it can support the neurodivergent members of your family or team, right alongside everyone else:
- One thing at a time — Deep Work Mode goes fullscreen on a single task, with a Pomodoro/count-up timer and a gentle "still working?" nudge. No tab, no feed, no next thing.
- The app decides what's next, not you — Today/Overdue views and the Morning Brief and Priority Focus agents surface what matters, so a stacked day doesn't start with a blank-page decision.
- Falling behind isn't a wall — the Reschedule Advisor proposes new dates for overdue tasks with one tap. Nothing is auto-applied; every agent suggestion waits for your yes.
- Calm by default — dismissable insight cards, no popups, no red-badge anxiety.
- Structure you can lean on — Routines auto-reset daily and lay the day out on a Day Arc, so the scaffolding is there without you rebuilding it each morning.
- Reusable named timers — save "Chicken Defrost", "Leave in 10", "Laundry" once and reuse them — small help for time-blindness.
- Regulation tools built in — guided Breathe (box / 4-7-8 / calm), Meditate, and Ambient Sound on the Apps page, each with an optional before/after mood check so you can see a session helped.
- A shared brain — Notebooks and Lists hold what working memory drops.
- Motion-sensitive by choice, not by default — animation (confetti, weather icons, widget jiggle, alarm shake, breathing pulse) can overwhelm rather than delight if you're sensitive to on-screen motion. A master Reduce Motion switch in Settings turns it all off at once, defaults on if your OS already asks for reduced motion, with per-effect toggles if you just want to kill one thing.
paperr automatically adapts to whatever device opens it — no separate apps to install.
| 🖥️ Desktop | 📱 Phone | 📺 Tablet |
|---|---|---|
| Sidebar navigation, dense multi-column layouts, and a slide-out dotAi chat drawer. | Bottom-nav, touch-first shell built around the customizable widget Home board. | The full app in a touch-optimized nav-rail shell — a wall-mountable dashboard for the kitchen or office, sized for a mounted display. |
The mode is auto-detected from screen size and touch support on first load, remembered from then on, and can be switched anytime from an in-app button (e.g. the desktop-icon button in the tablet header).
💡 Give an old tablet a second life. Since paperr just runs in a browser on your LAN, any tablet you already own — an old iPad, a Fire tablet, an Android slate — can be propped up or mounted as an always-on shared screen for the whole family or team, no app install required.
- Node.js 22.5+ (the server uses the built-in
node:sqlitemodule, which needs 22.5 or later) - (Optional) Python 3 — lets
npm installauto-provision paperr's own bundled AI server. Skipped safely if Python isn't found; add it later and re-runnpm installinserver/to enable it. - (Optional) Ollama or LM Studio — external local LLM providers, if you'd rather point dotAi at one of those instead of the bundled server
npx paperr # fetches paperr into ./paperr — add a folder name for a custom location
cd paperr
npm run install:all # installs root + server + client dependencies, creates server/.env
npm run dev # starts API (:3000) and client (:5173) togethergit clone https://github.com/biswasprateek/paperr.git
cd paperr
npm run install:all # installs root + server + client dependencies, creates server/.env
npm run dev # starts API (:3000) and client (:5173) togetherBoth land in the same place. The npm package is only a thin fetcher — paperr is a self-hosted app, not a library, so there's nothing to
npm installinto an existing project.
🔐
npm run install:allcreatesserver/.envfrom.env.example(if it doesn't already exist) and auto-generates real random values forJWT_SECRET/JWT_REFRESH_SECRET— no manual copy/paste step needed. It never overwrites secrets you've already set. To (re-)run just that step:npm run setup:env.
🤖 Bundled AI Server, installs itself:
npm run install:all(which runsnpm installinsideserver/) automatically provisions paperr's own local AI model vialitert-lm— see Why paperr for what it runs on. It looks for a Python interpreter on your machine to build a project-local virtual environment (server/ai/litert/venv):
Platform Looks for Windows py -3, thenpythonmacOS python3, thenpythonLinux python3, thenpythonIf no Python interpreter is found, this step is skipped with a warning and the rest of the install continues normally — install Python 3 and re-run
npm installinserver/to enable it later, or use Ollama/LM Studio instead (see Configuration below). Once installed, manage the server (start/stop, live memory usage, auto-offload when idle, model picker, downloading additional models) from Settings → paperr AI Server.
On first launch, the Setup Wizard walks you through creating your first space and admin account.
🌱 Want it pre-populated instead?
node scripts/seed-demo.cjscreates a fully-populated demo space ("The Miller Family", 4 members, realistic tasks/projects/notes/routines across every section) — log in as any of them with passwordpaperrdemo1(maya is admin). Re-running it wipes and recreates just that demo space.
Windows — one click:
./start-paperr.ps1 # builds the client, then serves everything on :3000
./start-paperr.ps1 -NoBuild # skip rebuild if the client is already currentmacOS / Linux — one click:
./start-paperr.shAny platform:
npm run build # bundles the React app
npm start # Express serves the API + app on port 3000Then open http://<your-machine-ip>:3000 from any device on your network.
paperr binds to 0.0.0.0, so it's reachable from every device on your LAN. Find your local IP (ipconfig on Windows, ifconfig/ip a on Linux/macOS) and open:
http://192.168.x.x:3000
💡 Tip: Set a static IP or DHCP reservation for the host machine so the URL never changes, and add a desktop/taskbar shortcut to
start-paperr.ps1for one-click launch.
server/.env is created automatically by npm run install:all (see Quick Start above), with JWT_SECRET / JWT_REFRESH_SECRET auto-generated. Edit server/.env directly to adjust anything else:
| Variable | Default | Description |
|---|---|---|
PORT |
3000 |
Server port |
JWT_SECRET |
(auto-generated) | Signs access tokens |
JWT_REFRESH_SECRET |
(auto-generated) | Signs refresh tokens |
DB_PATH |
./data/databases/paperr.db |
SQLite database file path |
UPLOADS_PATH |
./uploads |
File attachment storage |
LLM_BASE_URL |
http://localhost:11434 |
Ollama / LM Studio base URL |
LLM_MODEL |
llama3 |
Local model name for dotAi |
HTTPS_ENABLED |
false |
Enable TLS on the LAN (optional) |
paperr is built private-first, with real auth baked in:
- JWT access + refresh tokens stored in HTTP-only cookies
- bcrypt-hashed passwords (
bcryptjs) and quick-login PINs - Brute-force protection — accounts lock after repeated failed attempts
- Role-based access — admin vs. member, enforced server-side
- Session management and a full audit log of sensitive actions
| Layer | Technology |
|---|---|
| Runtime | Node.js 22.5+ (needs the built-innode:sqlite module) |
| API | Express.js |
| Database | SQLite via Node's built-innode:sqlite (single file, zero config, no native build step) |
| Auth | JWT (HTTP-only cookies) + bcrypt (bcryptjs) |
| Real-time | Socket.io |
| Scheduling | node-cron (backups, proactive AI agents) |
| Frontend | React 18 + Vite + React Router v6 |
| State | Zustand + React Query |
| Styling | Tailwind CSS |
| Editors / Charts / DnD / Dates | @uiw/react-md-editor · Recharts · @dnd-kit · date-fns |
| Font / Icons | Plus Jakarta Sans · Material Symbols ·emoji-picker-react |
| AI | Any OpenAI-compatible local LLM (Ollama, LM Studio, llama.cpp) + a bundledlitert-lm server paperr installs and supervises itself (Windows / macOS / Linux) |
| Images | sharp — downscales/recompresses curated Frame art server-side to match browser-uploaded photos |
paperr/
├── server/ Express API — routes, auth, SQLite, AI tools & agents
│ ├── routes/ tasks, lists, projects, calendar, notes, routines, spaces, agents, backups, …
│ ├── ai/ dotAi's LLM client, tool definitions, handler, scheduler, built-in agents, bundled AI server supervisor
│ ├── services/ backups, notifications, and other cross-route business logic
│ ├── data/ hand-curated starter content (Frame photo sets, Good Thoughts collections)
│ ├── scripts/ postinstall setup (auto-provisions the bundled AI server's Python venv)
│ └── db/ schema.sql + node:sqlite setup
├── client/ React + Vite app
│ └── src/
│ ├── pages/ Tasks, Projects, Calendar, Notebooks, Routines, Frame, Agent Hub, …
│ ├── modes/ Desktop / Phone / Tablet layouts
│ ├── components/ Shared UI — task form, Deep Work overlay, backups, agents, Good Thoughts, SpacePicker, …
│ ├── store/ Zustand stores (focus, deep work, clock, UI, …)
│ └── widgets/ Home board & Hub board widgets (Time Progress, AI Agent, Frame, Good Thoughts, …)
├── start-paperr.ps1 One-click Windows launcher
└── start-paperr.sh One-click macOS/Linux launcher
Bug reports, feature ideas, and PRs are welcome. See CONTRIBUTING.md for setup, expectations, and how to file a good bug report.
Built with ❤️ in NYC.
Apache 2.0 licensed. ⭐ this repo if paperr replaced a subscription for you.