ActivityWatch - Open-source time tracker

A lighter faster ActivityWatch with Tauri: now available for testing

Tue, 06 Oct 2026 12:00:00 +0000

The ActivityWatch v0.14.0 release ships a new desktop app: aw-tauri, a lighter, faster cross-platform repackaging of ActivityWatch. As the name implies the project is built with Tauri, a relatively new Rust-based toolkit that enables easy development of small, fast, and secure applications with a great developer experience.

We first announced this work in 2024, when it was little more than a prototype. Since then it has grown into something you can install on Windows, macOS (Apple Silicon and Intel) and Linux, and v0.14.0 includes it for every one of those platforms. This post is the story so far: why we did it, what you get, and how you can try it today.

The Activity view inside aw-tauri's own window (v0.14.0b8, Linux; demo data). No browser tab needed.

Why Tauri

Tauri apps are lightweight, memory efficient and secure by design. Tauri does not ship a renderer but uses the platform native renderer via WebViews. This simple design choice makes the app size compact and memory efficient during runtime, as compared to electron apps. Tauri apps are secure, only interacting with the host systems through Tauri APIs.

That last part holds if you build an app the “Tauri way”. ActivityWatch also runs an HTTP server (aw-server-rust, on localhost:5600), so we still have to take precautions to not expose sensitive information to other devices on the network or to websites, just like before.

What you get

aw-tauri aims to replace the functionality that is currently implemented in aw-qt and aw-notify. Creating a new Rust-powered entrypoint for running ActivityWatch and its modules. With aw-server-rust as the default server, it leads to a reliable and highly performant application. Reliability and performance is very important for apps like ActivityWatch, which always run in the background and collect data that is otherwise lost.

aw-tauri is designed to be a drop-in replacement for aw-qt, and the main application that appears as a tray icon and manages the server and the watchers:

It houses its own webview, no need to visit http://localhost:5600 in your browser anymore.
Watchers can be started and stopped right from the tray icon, just like in aw-qt. Crashed modules are restarted automatically, with a backoff and a limit on retries.
The server runs inside the app (aw-server-rust), so there is no separate server process to manage.
Autostart on login is built into the app and can be toggled from the tray menu, no need to set up systemd services for Linux users.
Notifications: aw-notify has since been rewritten in Rust, and aw-tauri runs it and shows its output as native desktop notifications, enabling configurable usage notifications (such as goals or alerts).
aw-sync is included in the bundles, but sync is still in preview: it is opt-in and not switched on by default.
It can update itself: from v0.14.0 the built-in updater (wired up to GitHub Releases) updates the app on macOS and Linux. On Windows, install new versions manually for now.

The watchers themselves are still mostly the same ones as before: the Python aw-watcher-afk and aw-watcher-window for cross-platform compatibility. There is no cross-platform Rust watcher, yet. On Linux the bundle also includes awatcher, a Rust watcher that covers window and AFK tracking on both X11 and Wayland, which is why the v0.14.0 release notes list native Wayland support for the Tauri build.

aw-tauri does not replace the classic aw-qt build in v0.14.0: both are published side by side, and the release notes still label the Tauri distribution as experimental. We would like to hear how it works for you before that changes.

Developer Experience

aw-tauri is built with aw-server-rust serving the backend together with aw-webui. We have taken care to keep the codebase lean and clean, such that it takes very little time to get acquainted with the codebase.

In our current build process for Python modules like aw-qt, we rely heavily on PyInstaller for building into binaries. This has been an enduring source of problems, and many developer weeks (if not months) have been spent trying to work out all the issues over the years (especially macOS support, because of the need to codesign a very messy .app bundle).

With Tauri, they have handled most of the heavy lifting, and make it easy to produce working binaries for all target platforms. Much of this is simply due to Rust (avoids PyInstaller and its complexity), but also the added tooling for codesigning, and producing suitable bundles for each platform: on Linux you get a lightweight .AppImage, on Windows an installer, and .app on macOS.

In v0.14.0, the Windows installer is an .exe setup, macOS gets a .dmg for both Apple Silicon and Intel, and Linux gets .AppImage, .deb, .rpm and .zip for x86_64 and arm64.

The Tauri bundles are also smaller to download. Here is the size of the installers attached to the v0.14.0b8 release, for the classic build and the Tauri build:

Installer	Classic (`aw-qt`)	Tauri
Windows `.exe`	81 MB	43 MB
macOS Apple Silicon `.dmg`	64 MB	53 MB
macOS Intel `.dmg`	70 MB	56 MB

These are download sizes read from the release assets, not a measurement of runtime memory use. We have not benchmarked memory yet.

Try it today

v0.14.0 includes aw-tauri builds for all platforms. Find them under “Tauri distribution” in the v0.14.0 release notes and assets. (The downloads page links the classic aw-qt build by default.)

Windows: the .exe installer.
macOS: the .dmg, for either Apple Silicon or Intel.
Linux: the standalone .AppImage, .deb and .rpm are self-contained with awatcher (window and AFK tracking on X11 and Wayland) plus aw-sync. The .zip includes those modules too, along with the Python watchers and move-to-aw-modules.sh for installing optional modules in ~/aw-modules/.

A few caveats, since this is still testing:

Back up your data before upgrading. The v0.14 server performs a one-time database index migration on first start. It does not intentionally change or delete events, but on a large database it can take a while, so leave ActivityWatch running until the dashboard shows up.
Don’t run aw-qt and aw-tauri at the same time: they both want port 5600.
Found a bug? Please open an issue or tell us in the release discussion.

Who this is for

ActivityWatch has passed 1 million downloads, and the desktop app is where most of our users start. Making it simpler to build, sign and ship is how we get fixes and improvements out to them faster.

Cumulative and weekly desktop downloads from GitHub Releases. Live chart on the stats page.

Help us finish it

aw-tauri is still in active development and contributions are welcome and encouraged! More eyes on the code will be beneficial to the project.

Check out the README to see the status of development, and where you can help out. Open items include a settings UI, downloading watchers on demand, the updater, and code signing (see the issues).

We are hopeful that it will help solve many of our remaining challenges, and are excited to see it help shape the future of ActivityWatch.

Thanks to Brian Vuku (0xbrayo), the main author of aw-tauri, who wrote the original version of this post.

Syncing your ActivityWatch devices: what works today, what's next

Tue, 06 Oct 2026 10:00:00 +0000

You can now copy activity data between your phone and your computers without an ActivityWatch account. ActivityWatch has no sync server of its own; whether the copies stay on machines you control depends on the folder tool you pick. Sync and multi-device queries have reached the point where the basics work and Erik uses them on his own devices. They are not yet click-and-play: some setup is manual, and a few pieces — including the combined Activity view — are merged but not yet in a release. This post covers what works, what needs care, and what is still ahead.

How it works

ActivityWatch has no sync server. Each device keeps its own data locally, as it always has. aw-sync adds one step: it copies a device’s buckets into a sync folder, and it reads the other devices’ copies out of that folder.

The sync data flow as it works in v0.14.0. Names and layout will change with the v2 redesign below. On a small screen, scroll sideways or open the diagram full-size.

The moving parts:

Push. Every device writes only into files it owns: a staging database at {hostname}/{device_id}/test.db inside the sync folder (default ~/ActivityWatchSync on desktop). Nothing else writes to it.
Transport is your choice. Share the folder with a tool you already use. Syncthing, rsync, or a self-hosted share keep copies on machines you control. Dropbox or Google Drive also work, but then the folder copies go through that provider’s servers. aw-sync does not move files between machines; that is deliberate, so you pick where your data travels.
Pull. A desktop opens the other devices’ databases read-only and imports their events into its own aw-server as buckets named like aw-watcher-android-synced-from-<host>.
View. Once imported, the web UI can query local and imported buckets together. In v0.14.0 the Activity view can merge any set of devices: pick them in its device selector.

What works today

Android pushes. In Android 0.14, Sync Settings lets you switch sync on (it is off by default) and pick a folder through Android’s storage picker. Since 0.14.2 it also shows when the next sync is scheduled and what each pass did.
Desktop pulls. aw-sync sync does one pull and push pass. aw-sync status shows which peers it found in the folder and what the last pass did.
Multi-device queries. They exist in the web UI and have been fixed several times in the last few weeks, including for phones that have no AFK data. v0.14.0 ships an All devices / pick-devices selector in the Activity view (aw-webui#1004).

What needs care today

We would rather tell you than have you find out.

It’s early. Sync is opt-in and not started by default, and setup is manual.
Android does not pull yet. The phone sends its data out but does not import your desktops’ data (aw-android#291). The app’s own web UI therefore shows only the phone’s data.
You trigger the desktop pull. The background daemon only pushes by default in v0.14.0; set pull = true under [daemon] in aw-sync’s config.toml to have it pull too. Running aw-sync sync yourself or from a timer always does both. The sync docs cover setup.
The multi-device view covers apps and categories. Browser and stopwatch data stay single-device in that view.
Some limits are built in. Settings are not synced, deleting an event does not propagate, and every device mirrors events from every other, so many devices means many copies and more disk. Early versions could import duplicate events; aw-sync dedupe --dry-run reports them.
Imports are not read-only. Local edits to an imported bucket are silently lost (aw-server-rust#694). Edit on the device that owns the data.

What’s next

A leaner staging area (sync v2). The staging database is the wrong shape for sync. Version 2 is designed to replace it with immutable, compressed segment files plus a manifest, with the goal of more predictable imports and less space used. The first piece, a segment writer, has merged behind a feature flag and is not used by any command yet. Design and progress are tracked in activitywatch#1445.
Android pull. It is waiting on v2, so that mirroring every device’s data does not fill up phone storage.
Finishing multi-device queries. Getting the web UI selector released, and making sure the clients, phone included, actually get multi-device data.

We are not putting dates on these.

Try it, and tell us what breaks

Pick a folder and share it between your devices with a tool you trust. If you want copies to stay off third-party servers, use Syncthing, rsync, or a share you host.
On the phone, turn on Sync in the navigation drawer and choose that folder.
On the desktop, install v0.14.0 and run aw-sync sync, then check aw-sync status.
Confirm the import from aw-sync status (it lists discovered peers) and from the raw buckets list or Timeline, which should show ...-synced-from-<host> buckets. Then pick your devices in the Activity view’s device selector to see them together.

Expect the first pull to import everything already in the folder, which can take a while for a phone with months of history.

If something goes wrong, open an issue and include the aw-sync status output after redacting identifying peer names and paths, or ask on the forum. Reports from real setups are what makes this better.

ActivityWatch v0.14.0: faster, steadier, synced, and a new app

Tue, 06 Oct 2026 10:00:00 +0000

ActivityWatch v0.14.0 is out. It’s been almost two years since v0.13.2, and it’s the biggest release in a long time: about 1,000 changes across 13 repositories, from more than 30 contributors, tested over eight betas.

A lot has happened since the last stable release. ActivityWatch passed 1 million downloads along the way, and it’s now downloaded over 10,000 times a week. Thank you!

The short version: it’s much faster once you have years of data, a lot steadier, your devices can finally share their data with each other, the Android app has caught up, and there’s a new desktop app built with Tauri. The longer version is below.

Download v0.14.0 · Full changelog

Your devices, together

Multi-device support is the most-requested feature in our user survey (1,714 responses). You use a laptop, a desktop and a phone, and you want one picture of your time, not three.

v0.14.0 is the release where that comes together. Sync is local-first: there is no ActivityWatch server and no account. aw-sync copies each device’s data into a folder, and you move that folder between devices with whatever you already trust: Syncthing, Dropbox, a network drive. Your data never touches a server of ours. Each device only ever writes its own files and opens the others read-only, so one device can’t corrupt another’s data.

Once your devices are syncing, the Activity view can show any set of them merged into one view: your desktop and laptop together, or everything including your phone. Pick the devices in the device selector.

We spent the last weeks of this cycle running sync on our own machines and phones and fixing what broke: duplicate events after a resume, one bad device aborting a whole sync, phone data not being picked up, and an aw-sync status command that now tells you what each device has synced.

It’s honest to say it’s still rough. Every device keeps a full copy of every other device’s data, so it uses more disk than it should. Settings and deletions don’t sync. Setup is manual, and sync isn’t started for you: turn it on in the app’s modules menu, or run aw-sync. A much more compact format is already in the works, and making multi-device simple is the main goal for the next releases. If you run ActivityWatch on more than one device, try it and tell us what you hit.

Android caught up

Last month we released ActivityWatch for Android 0.14, the first stable Android release since 2023. User-perceived crashes went from about 7.8% to under 1%, and it added alerts, a home-screen widget, Firefox and Chrome tracking, and CSV and JSON export.

It also ships the same sync as the desktop: pick a folder on your phone, sync it like any other, and your phone’s time shows up next to your desktop’s in the Activity view.

Much faster on large databases

If you have used ActivityWatch for a few years, you have probably watched the dashboard get slower. We measured v0.13.2 against v0.14 on a real 9-year database (9.2 million events, 1.8 GB), with each version’s own dashboard queries. The new dashboard loads long ranges one day at a time, so here are both servers on that same workload, a Year of daily queries:

Year view, 365 daily queries	v0.13.2	v0.14 first load	v0.14 repeat load
Classic app (Python server)	120 s	41 s	0.6 s
Tauri app (Rust server)	139 s	32 s	0.2 s

Loading day by day is also what lets the dashboard show progress, draw per-day charts, and cache days that are over, which is why coming back to a view is near instant. (v0.13.2 asked for a whole Year in one request instead: on the Python server that took two minutes and hit the 30-second timeout, on the Rust server 19 seconds.)

Under the hood there are new database indexes, SQLite’s WAL mode, cached category rules (categorizing a year of window titles went from 53 seconds to 10) and exports that stream instead of loading everything into memory. Day, week and month views benefit too: they no longer time out on big databases.

Steadier

ActivityWatch runs in the background all day, and when it breaks you lose time you can’t get back. A lot of this release went into making that rarer:

Watchers that crash now restart on their own, in both apps, with crash logs
On macOS, the window watcher no longer crashes on unusual window titles, and a leak of up to ~359 MB of memory a day on busy machines is fixed
Locking your screen now counts as AFK on macOS, gamepads count as activity on Linux, and a Windows bug after 49.7 days of uptime is gone
A damaged database recovers on startup, and imports merge into existing data instead of failing
The Python and Rust servers now give the same query results, checked by a new test suite (how the two servers compare)

A new app, built with Tauri

v0.14.0 ships a new desktop app, aw-tauri, alongside the classic one. It has the same features and is lighter: it has its own window, runs the Rust server inside the app, is smaller to download, supports Wayland on Linux out of the box, and updates itself on macOS and Linux. It’s less battle-tested so far, so the classic app ships too and stays the default.

Brian, who built most of it, wrote about why and how: A lighter, faster ActivityWatch with Tauri.

macOS builds of both apps are now signed and notarized for Apple Silicon and Intel, so they open without Gatekeeper workarounds, and browser URLs are tracked properly again.

A refreshed web UI, and more control

Timeline: swimlanes, filters for AFK time, categories and duration, and “merge by app” to calm down busy days
Custom date ranges, and a Year view that is always available
Category sets: keep several sets of categorization rules and switch between them, plus rule priorities
A Work Time report and billable hours export by category
A privacy filter that drops sensitive window titles before they ever reach the database
An optional API key for the server, if you expose it beyond your own machine
ActivityWatch now speaks Swedish, German, Ukrainian, Russian and Chinese
Support for more browsers: Arc, Dia, Zen, Helium, Floorp and others

And for the AI crowd: one canonical query returns clean, categorized events, so an assistant can answer questions about your time without a custom integration. See ActivityWatch with agents and AI.

Upgrading

On the first start after upgrading, ActivityWatch updates its database indexes once. Your data isn’t changed. On a large database it takes up to a minute, and the dashboard may not connect until it’s done. Leave it running and reload. Notifications are now opt-in, so turn them back on in Settings if you used them.

Thank you

More than 30 people contributed to this release, and over a million of you have downloaded ActivityWatch. Many of the features above came from the community: @0xbrayo built aw-tauri; @BohdanRykush, @JhihJian and @NickWick13 made ActivityWatch speak Ukrainian, German, Russian, Chinese and Swedish; @BelKed rewrote the browser extension for Manifest V3; @2e3s’s awatcher brings Wayland support; @nerumo added timeline swimlanes; @rrogerc and @hawai-i got browser URLs working on macOS; and @FractalMachinist and @Lorite improved browser tracking and sync on Android. The release notes credit everyone else. Thank you all.

ActivityWatch has no ads, no venture funding and no data business. If it is useful to you, ActivityWatch Pro, our patronage subscription, is the most direct way to keep releases like this coming. It starts at $5 a month and unlocks nothing: everything stays free.

ActivityWatch Research Edition: a variant for your study

Tue, 29 Sep 2026 22:00:00 +0000

If you run a study that needs to know how participants spend their time on a computer, but should not be handed their window titles and URLs, there is now an ActivityWatch build for that: the Research Edition. Making a variant for your study is mostly a matter of supplying a category map, so if you are a researcher, get in touch and we can work out what fits. Read the docs for the full details.

Where it came from

Researchers want categorized time, not raw activity: “this much on communication, news and work tools”, without the research team ever seeing which page or document that was. The Research Edition grew out of academic studies that wanted exactly this, the current one at Lund University, and it follows an approach an earlier academic team had tried in a public fork of ActivityWatch.

What it does

The Research Edition is a separate build of ActivityWatch with a privacy filter switched on. As of the latest build, v0.14.0b5-research (a prerelease, for Windows, macOS and Linux):

The window watcher classifies each browser window into a study category, using the URL where available and the window title otherwise, then discards the title and URL before anything is written to disk. Anything the map does not match becomes excluded.
Non-browser windows lose their titles too.
Time away from the computer is recorded, like in regular ActivityWatch.
Data stays on the participant’s computer. At the end of the study they export one JSON file from the dashboard and upload it to the study team. The export also strips the computer’s name. There is no live telemetry and no study server to run.

What the Research Edition keeps, and where the filtering happens.

The Research Edition docs page has the complete list of what is and is not stored.

Two honest caveats. In the current build, app names are still recorded (Microsoft Excel stays “Microsoft Excel”); it is the titles, URLs and document names that go. The filter can also replace app names with categories through a config option, so a variant that needs that is a configuration choice, not new code. The filter is only as good as the category map, which is why the map belongs to the study. And it covers the bundled window watcher: if a participant connects other watchers to the Research Edition, their data ends up in the same export, so studies should tell participants not to.

It also keeps participants’ own data separate from the study’s. The Research Edition has its own app identity, its own data folder and its own port (5667 instead of 5600), so a participant who already uses ActivityWatch can run both side by side without mixing the two. That lets an ethics approval say “study data is kept apart from the participant’s own”, and it can be checked, not just promised.

How much data

Small: for a default install, roughly 1-5 MB per participant per week of collection, as an estimate. It is one JSON file, exported by the participant. Extra watchers add more, so if you need a firm number for a data-protection review, measure a pilot week.

Making a variant for your study

The design is meant to be picked up by the next researcher without starting from scratch. A variant is mostly:

Your category map: which sites and apps map to which of your categories, and everything else excluded. This is the part only you can define.
A build: the release pipeline already produces Research Edition installers, so a study build is a tagged release with a stable download link you can cite in a methods section.
A participant guide and export step to match your protocol. The docs have participant instructions you can link to or adapt. They cover Windows and macOS; a Linux version can be written if a study needs one. An Android version exists for the Android build, which is not released yet.

New studies tend to want what the previous one wanted: categorized time, no raw titles or URLs, an easy export, and eventually mobile. Ethics committees decide for themselves, but the design is built around data minimisation: the bundled window watcher drops titles and URLs before storage, nothing is sent live, and the participant exports a single file they can inspect. Bring your committee’s requirements and we can adjust the build to them.

Mobile

Android: the Android app has a research build flavor with its own app id and port, so it can be installed next to the regular app without touching its data. This covers the separation only. The category filter you get on desktop is not on Android yet, so an Android Research Edition is in progress, not shipped.
iPhone and iPad: there is no iOS app. Instead, aw-import-screentime imports Apple’s Screen Time data into ActivityWatch. It works for any participant who also uses a Mac, with “Share Across Devices” turned on for both devices under the same Apple account, because Screen Time syncs between them. It needs Full Disk Access on the Mac. It is a standalone tool today, and the study filter does not apply to that data path yet.

Get in touch

Everything for researchers (the Research Edition, participant instructions for desktop and Android, and how to cite ActivityWatch) is collected in the ActivityWatch in research section of the docs.

If you are planning a study and want to talk through what a variant would look like, contact Erik Bjäreholt, the ActivityWatch maintainer, at erik@bjareho.lt (also on his GitHub profile). The Research Edition docs describe what a variant consists of. Tell us which categories you need, which platforms your participants use, and what your ethics process requires.

ActivityWatch has passed 1 million downloads

Tue, 29 Sep 2026 10:05:00 +0000

ActivityWatch has passed 1 million downloads. The desktop app’s GitHub Releases alone now count over 1.18 million, and that doesn’t include package managers, distro repositories, Google Play, or F-Droid. We didn’t mark the moment when it happened, so this is us catching up.

Cumulative and weekly desktop downloads from GitHub Releases. Live chart on the stats page, along with users, stars, and Android numbers.

The weekly rate has roughly quadrupled since 2023 and is now over 10,000 downloads a week.

ActivityWatch is free, open source, and stores your data on your own machine. It has been built mostly by volunteers, so a number like this means a lot to us. Thank you to every user, and to every contributor who has written code, watchers, translations, docs, or bug reports.

What’s next

The stable desktop release, v0.14, is what we’re preparing now. On mobile, ActivityWatch for Android 0.14 just went stable, with sync, alerts, and a large batch of crash fixes.

Help us keep going

Star or watch ActivityWatch/activitywatch.
Tell us what’s missing in the forum or on GitHub.
Sponsor the project if it’s useful to you.

ActivityWatch for Android 0.14 is stable: sync, alerts, and a lot fewer crashes

Tue, 29 Sep 2026 10:00:00 +0000

ActivityWatch for Android 0.14 is out of beta. It’s the first stable Android release since v0.12.1 in 2023, and the biggest one the app has had: 30 new features, 100 bug fixes, and contributions from 32 people. Since v0.14.0 (Sept 13), two follow-ups, v0.14.1 and v0.14.2, have gone out to tighten things up.

The first Android release in nearly three years is mostly a stability release, so we’ll start there.

Crashes and performance

The numbers from Google Play show it. User-perceived crash rate averaged about 7.8% from January to late August 2026 and has been about 0.9% in September, as the 0.14 builds reached Google Play.

Android user-perceived crash rate, from Google Play vitals. Live chart on the stats page.

What changed:

Startup hangs. A slow datastore used to turn app start into a hang. Startup paths are now hardened (#262).
StackOverflowError in the watcher. Accessibility tree traversal is now bounded, so devices with deep view hierarchies no longer blow the stack (#269).
Off-thread initialization. The Rust interface is constructed off the calling thread instead of blocking it (#277).
Rotation. Rotating your phone no longer reloads the WebView and throws away the view you were on (#271).
Onboarding. MainActivity now relaunches correctly once onboarding finishes (#298).

Sync with your desktop

Your phone’s data can now flow into the same reports as your desktop. Turn on Sync in the navigation drawer and choose a folder that Syncthing, Dropbox, Drive, or another file-sync tool shares with your desktop’s ~/ActivityWatchSync folder. Then run aw-sync sync on the desktop to import the phone’s buckets. aw-sync does not transport the files itself. Since desktop v0.14.0, the background sync daemon can also pull the phone’s data on every pass: set pull = true under [daemon] in aw-sync’s config.toml. See the sync docs for setup.

Updated 2026-10-06: this post originally said the desktop daemon couldn’t read the Android folder layout. That was fixed in desktop v0.14.0.

As of 0.14.2 the app also shows when the next sync is scheduled and what each sync pass actually did, so “why isn’t my phone data showing up?” now has an answer in the UI.

Activity alerts (via aw-notify): get notified when you’ve spent too long in a category.
Home screen widget with category time, refreshed every five minutes.
Firefox and Chrome browser tracking.
Export to CSV/JSON through the share sheet now works reliably.
A Material You monochrome icon for Android 12+.

Get it

Update via Google Play, or grab the APK from the GitHub release.
Live Android stats (installs, rating, crash rate): activitywatch.net/stats
Full changelog: v0.14.0, v0.14.1, v0.14.2.

If something still breaks, open an issue. Thanks to everyone who reported problems during the beta, and to the contributors who fixed them.

ActivityWatch vs RescueTime: Why I Built a Free Open-Source Alternative

Mon, 02 Mar 2026 08:37:00 +0000

The Story Behind ActivityWatch

I started building ActivityWatch because I was frustrated with RescueTime. As someone who wanted to understand how I spent my time on the computer, I had been a RescueTime user for years. But over time, the problems kept piling up:

My data wasn’t mine. RescueTime stores everything on their servers. I had no way to self-host or keep my data local. For something as personal as a complete log of everything I do on my computer, that felt wrong.
It wasn’t open source. When I found bugs — like Xbox controller input being tracked incorrectly — I couldn’t fix them myself. I filed reports, but they went nowhere.
Linux support was an afterthought. As a Linux user, I was a second-class citizen. The Linux client was always behind, and eventually RescueTime dropped Linux support entirely.
“Pay or lose your history.” The free tier only showed the last 3 months. Want to see your all-time data? Pay up. It felt like they were holding my own data hostage.
No extensibility. I couldn’t add custom watchers, build integrations, or extend the tracking in any meaningful way. What you got was what you got.

So I built ActivityWatch — a free, open-source, privacy-first time tracker that keeps your data on your device and lets you extend it however you want.

Feature Comparison

Here’s how ActivityWatch compares to RescueTime in 2026:

Feature	ActivityWatch	RescueTime
Price	Free and open-source	$12/month (Premium)
Data storage	Local (your device)	Cloud (their servers)
Linux support	Full support	Discontinued
Open source	Yes (MPL-2.0)	No
Custom watchers	Yes — write your own	No
Browser tracking	Chrome, Firefox, Edge, Opera	Chrome, Firefox, Edge
Offline tracking	Yes	Limited
API access	Full local REST API	Limited API
Self-hosting	Built-in (it’s local-first)	Not available
Data export	Full export anytime	Limited in free tier
Distraction blocking	Via third-party tools	Built-in (Premium)
Team features	Coming soon	Available (Premium)
Mobile tracking	Android	Android, iOS

Where ActivityWatch Wins

Privacy and Data Ownership

This is the fundamental difference. ActivityWatch stores all your data locally on your device. No cloud accounts, no data uploads, no risk of a company going under and taking your data with it. You own your data, period.

Extensibility

ActivityWatch has a modular architecture. The core is a local server that stores events from “watchers” — small programs that observe different things. You can write your own watchers for anything: IDE activity, music listening, exercise, or whatever you want to track.

The community has built watchers for:

VS Code and JetBrains IDEs
Spotify and media players
VR headsets
Custom hardware

Cross-Platform (Including Linux)

ActivityWatch runs on Windows, macOS, and Linux. When RescueTime dropped Linux in 2024, many users switched to ActivityWatch. We’re committed to supporting all major platforms — it’s a core part of who we are.

No Paywalls on Your Own Data

All your data is always accessible. There’s no “upgrade to see your history” or “pay to unlock analytics.” Every feature is available to every user.

Where RescueTime Still Has an Edge

To be fair, RescueTime does some things well:

Built-in distraction blocking — FocusTime lets you block distracting websites. ActivityWatch focuses on tracking; for blocking, you can use tools like Cold Turkey or LeechBlock.
Team management — RescueTime has built-in team dashboards and management features. ActivityWatch is primarily designed for individuals (though team features are on our roadmap).
iOS support — RescueTime has an iOS app. ActivityWatch currently only supports Android for mobile.
Polished onboarding — RescueTime has a smoother out-of-box experience for non-technical users. We’re actively working on improving our onboarding.

Migrating from RescueTime

Switching to ActivityWatch is straightforward:

Download ActivityWatch from activitywatch.net/downloads
Install and run — it starts tracking immediately
Install the browser extension for Chrome, Firefox, or Edge
Check the dashboard at localhost:5600 to see your data

Your RescueTime data can be exported and potentially imported — check our documentation for details on data import options.

What Users Say

People who switched from RescueTime consistently highlight:

“I switched when RescueTime dropped Linux. ActivityWatch is better in every way — I can actually see all my data without paying.”

“As a developer, the extensibility is what sold me. I wrote a custom watcher for my terminal in an afternoon.”

“The privacy aspect is huge. I don’t want a company having a complete log of everything I do on my computer.”

Try ActivityWatch

ActivityWatch is free, open-source, and takes about two minutes to set up. Download it at activitywatch.net/downloads and see for yourself.

If you’re coming from RescueTime, you’ll find that ActivityWatch tracks everything RescueTime did — and more — while keeping your data private and giving you full control.

Links:

Announcing Winners of Screenshot Competition!

Mon, 30 Oct 2023 09:56:00 +0000

We’re excited to announce the winners of the ActivityWatch Screenshot Competition! We received some great submissions, and we’d like to thank everyone who participated. Your contributions will help us showcase ActivityWatch to the world!

Submissions were accepted in the GitHub discussion announcing the competition, and the #screenshots channel on our Discord.

We’d like to apologize for the delay in announcing the winners. We are grateful for everyone’s patience and enthusiastic participation in this competition.

You can find selected screenshots from the winners on our screenshots page.

The Winners

🥇 1st place: Erik (@thelifeimliving)

For the sheer number of screenshots submitted, and of very high resolution, Erik takes the top spot. Your contribution will greatly help us in showcasing the versatility of ActivityWatch!

🥈 2nd place: Hernan (@realhyperchango)

Hernan showed us how category rules can be set up to get productivity classification instead of just activity classification. This is incredibly useful for users focused on time management.

🥉 3rd place: @BelKed and @maholli

For third place we have a tie between @BelKed and @maholli.

@BelKed showed us an excellent dark mode screenshot. It shows how elegant ActivityWatch can look in different themes.

@maholli showed us a whole month of activity! Good luck with your thesis!

Prizes

All winners will have their screenshots featured on the ActivityWatch website.

We were supposed to have additional prizes, which we still aim to give out, but we have yet to order the swag we planned to ship. However, winners can expect to get swag for free as soon as we order it!

We promise the winners will one day get ActivityWatch swag, including:

mug
T-shirt
stickers

We’ll reach out to the winners to distribute the prizes once we have them. (sorry if it takes a while!)

Next Steps

A big thank you to everyone who participated! Your contributions are invaluable.
The selected screenshots will be showcased on our website.

Further Discussion

For further discussion, a GitHub thread has been set up. Feel free to share your thoughts here.

The Future of Time Tracking: AI, Privacy, and Personalization

Wed, 30 Aug 2023 13:57:00 +0000

Introduction

In an increasingly digital world, time-tracking software like ActivityWatch plays a pivotal role in enhancing productivity and self-awareness. As we look to the future, three key elements stand out: Artificial Intelligence (AI), privacy, and personalization.

The Role of AI in Time Tracking

The Power of Data

While ActivityWatch itself doesn’t currently utilize AI, the data it collects serves as a fertile ground for AI-driven applications. The software logs a comprehensive array of your digital activities, from the websites you visit to the applications you use. This data can be transformed into actionable insights through machine learning algorithms. For example, natural language processing (NLP) could analyze your text-based activities to gauge your focus levels, while clustering algorithms could categorize your activities into productivity zones.

Long-Term Behavioral Trends

AI’s capability to sift through large datasets makes it invaluable for identifying long-term behavioral changes or trends. By applying time-series analysis on your ActivityWatch logs, you can gain insights into how your work habits, focus periods, and even leisure activities evolve over time. This can be particularly useful for self-improvement and for tailoring your work schedule to your natural rhythms.

Day Summaries and Note-Taking

Imagine an AI assistant integrated into ActivityWatch that can not only summarize your day but also take context-aware notes for you. Utilizing speech-to-text algorithms during meetings, it could generate summaries and action items automatically. When you’re working on a project, it could log your key milestones based on your activity data, thereby creating a daily journal without any manual input.

How It Could Work:

Speech-to-Text: During meetings, the AI assistant could transcribe and summarize key points.
Text Summarization Algorithms: At the end of the day, a summary of your activities could be generated, highlighting your most productive periods and areas for improvement.

In-Context Suggestions

Think of this feature as a modern, less intrusive version of Clippy. By leveraging real-time analytics and historical data, ActivityWatch could offer spontaneous, in-context suggestions. For instance, if the software detects that you’ve been working on a coding project for an extended period, it could suggest taking a break or even recommend a relevant Stack Overflow thread based on your recent queries.

Technical Details:

Contextual Bandit Algorithms: These could be used to offer suggestions that are most likely to be useful to you at any given moment.
Collaborative Filtering: By analyzing data from multiple users, the system could offer suggestions that have been beneficial to people with similar work patterns.

Integrating Brain-Computer Interfaces with Screentime Data: The Next Frontier in Time Tracking

The future of time tracking isn’t just about what you’re doing on your computer; it’s about understanding the cognitive processes that accompany those activities. Building on the pioneering research in my MSc thesis, which explored the classification of brain activity using electroencephalography (EEG) and automated time tracking, we see a new frontier emerging.

The Synergy of EEG and ActivityWatch

ActivityWatch provides a robust framework for tracking digital activities, while EEG devices offer a window into the cognitive state of the user. When combined, these two technologies can provide unprecedented insights into not just what you are doing, but how you are mentally engaging with those activities.

Technical Insights:

Data Fusion: By synchronizing EEG data with ActivityWatch logs, we can create a multi-modal dataset that captures both behavioral and cognitive aspects.
Machine Learning: Advanced classifiers based on Riemannian geometry, as demonstrated in my MSc thesis, can be employed to analyze this rich dataset.

Real-World Applications

The practical implications of this integration are vast:

Personalized Productivity: Imagine a system that knows when you’re most focused and automatically blocks distracting websites during those periods.
Mental Health Monitoring: By tracking cognitive states alongside digital activities, we can identify stressors and recommend timely interventions.
Enhanced Learning: For educational software, understanding the cognitive load during different tasks can help adapt the material in real-time to optimize learning.

Ethical and Privacy Considerations

While the potential is exciting, ethical considerations, particularly around privacy and consent, are paramount. All EEG data would be encrypted and stored locally, in line with ActivityWatch’s privacy-first philosophy.

Future Research and Development

As we look to the future, the integration of EEG devices and ActivityWatch opens up new avenues for research, particularly in the realms of cognitive science, human-computer interaction, and even preventive healthcare.

By marrying the behavioral data from ActivityWatch with the cognitive insights from EEG, we’re not just tracking time; we’re understanding it on a whole new level.

Full-Context Tracking: A Double-Edged Sword

Rewind.ai takes time tracking to the next level by offering full-context tracking, capturing not just screen data but also audio and microphone inputs. While this provides a rich dataset that can fuel advanced AI applications, it also introduces a set of challenges and ethical considerations. (Disclosure: I am an investor in Rewind.ai)

Opportunities for AI Applications

The depth of data collected by full-context tracking opens the door to a multitude of AI-driven applications:

Voice Analytics: By analyzing audio data, AI can gauge the sentiment and tone of meetings, providing insights into team dynamics.
Content Classification: Advanced machine learning models can automatically categorize screen content, offering a more nuanced understanding of user behavior.
Real-Time Feedback: With access to both screen and audio data, AI can offer real-time suggestions or alerts, such as flagging potential security risks or recommending breaks based on detected stress levels.

Privacy Concerns

While the wealth of data offers numerous possibilities, it also raises significant privacy concerns:

Data Encryption: Storing audio and screen data necessitates robust encryption methods to prevent unauthorized access.
User Consent: It’s crucial to obtain explicit user consent for capturing and analyzing such sensitive data.
Data Minimization: The principle of collecting only the data that is strictly necessary should be adhered to, in line with GDPR and other privacy regulations.

Resource Consumption

Full-context tracking is not just data-intensive but also resource-intensive:

Storage: The sheer volume of audio and screen data can quickly consume local storage, requiring efficient data compression algorithms.
Processing Power: Real-time analysis of such data can be CPU-intensive, potentially affecting system performance.

Rewind.ai has addressed this by leveraging it’s macOS-only nature to utilize Apple’s CoreML framework for on-device AI processing, allowing efficient data compression and real-time analysis with relatively low impact on overall system performance.

Balancing Act: AI, Privacy, and Resources

The challenge lies in striking a balance between leveraging AI’s capabilities and maintaining user privacy and system performance. One potential solution could be edge computing, where data is processed locally on the user’s device, minimizing data transfer and storage while still enabling real-time AI analytics.

The Importance of Privacy

The Privacy Imperative

In a world where data breaches are increasingly common, privacy is paramount. ActivityWatch addresses this by storing all your data locally, ensuring that you have complete control over your information.

The Resource Trade-Off

While full-context tracking offers richer data, it is much more resource-intensive and poses greater privacy risks compared to ActivityWatch’s local storage approach.

The Need for Personalization

Beyond One-Size-Fits-All

ActivityWatch allows for custom plugins and scripts, enabling you to tailor the software to your specific needs. This is crucial in a world where no two users are the same.

The Need for Researcher Tools

The Value of Data in Research

In the realm of behavioral and social sciences, the availability of accurate and comprehensive data is often the cornerstone of impactful research. Traditional methods of data collection, such as surveys and interviews, have their limitations in terms of scale and objectivity. This is where time-tracking tools like ActivityWatch can fill a significant gap.

Addressing the Lack of Data Collection Tooling

One of the challenges researchers face is the lack of specialized tooling for data collection that is both ethical and efficient. ActivityWatch addresses this void by offering a platform that is not only privacy-centric but also highly customizable. Researchers can use custom plugins and scripts to tailor data collection to the specific needs of their study, thereby enhancing the quality and relevance of the data collected.

ERC-Funded Project: A Case Study

ActivityWatch is currently being utilized in a 5-year project funded by the European Research Council. The project aims to delve into the intricacies of human behavior and time management. The use of ActivityWatch in such a high-profile research initiative underscores its value as a reliable tool for academic and scientific inquiry. Learn more about the ERC-funded project here.

WARN-D Research Project: The Broader Context

While not directly using ActivityWatch, the WARN-D research project aims to build a personalized early warning system for mental health by tracking stressors among students. This project highlights the broader societal value of time-tracking data and sets a precedent for how ActivityWatch could be employed in similar research endeavors. Learn more about the WARN-D project here.

By offering a robust, customizable, and privacy-focused platform, ActivityWatch is poised to become an invaluable asset in the toolkit of modern researchers.

Leading the Charge: ActivityWatch’s Role in Shaping the Future of Time Tracking

ActivityWatch is not just another time-tracking tool; it’s a platform that embodies the future of this technology. Situated at the crossroads of data-driven AI applications, stringent privacy measures, and extensive personalization capabilities, ActivityWatch is setting the standard for what time-tracking software can and should be.

By offering a platform that is both customizable and privacy-centric, ActivityWatch provides a unique solution that caters to a diverse user base—from individuals seeking to improve their productivity to researchers aiming to collect valuable data for scientific inquiry.

As we look to the future, ActivityWatch is committed to further enhancing its platform to meet the evolving needs of its users. Whether it’s integrating more advanced data analytics or expanding the range of custom plugins, the roadmap for ActivityWatch is geared towards innovation and user empowerment.

In a world where time is our most valuable resource, ActivityWatch offers a way to make the most of it. By focusing on the key pillars of AI, privacy, and personalization, ActivityWatch is not just keeping pace with the future of time tracking—it’s leading the way.

Conclusion

The future of time tracking is intricately tied to AI, privacy, and personalization. With its focus on these areas, ActivityWatch is well-positioned to lead the way.

If you are excited about this like we are, consider joining the Discord server. We’re always looking for new contributors and ideas!

And if you aren’t using ActivityWatch already, now is the perfect time to start collecting your screentime data!

Comparing ActivityWatch with RescueTime, ManicTime, and Apple ScreenTime

Wed, 30 Aug 2023 13:55:00 +0000

Introduction

In today’s digital landscape, effective time management is more critical than ever. While everyone has 24 hours in a day, how we utilize those hours significantly impacts our productivity and well-being. Time-tracking software has thus become a fundamental tool for individuals and organizations aiming to optimize their time use. This article will conduct a side-by-side comparison of four popular time-tracking applications: ActivityWatch, RescueTime, ManicTime, and Apple ScreenTime.

Feature Comparison

Here’s a table outlining the features of each software:

Feature Category	ActivityWatch	RescueTime	ManicTime	Apple ScreenTime
Tracking Capabilities	Granular tracking of applications, browser activity, custom event tracking. Extensible architecture.	Application and website tracking, limited offline tracking, distraction-blocking features.	Comprehensive computer usage tracking, supports offline activity tagging.	Limited to Apple ecosystem; tracks applications and website usage. No custom event tracking.
Reporting & Analytics	Rich dashboard with real-time and historical data. Custom reports possible due to open-source nature.	Detailed reports, limited customization.	Highly detailed reports, manual tagging and categorization.	Basic reporting on screen time and app categories, little customization.
Privacy	Local data storage by default. Robust privacy controls.	Data stored on their servers.	Local and cloud storage options.	Encrypted, on-device storage. Tightly integrated with Apple ecosystem.
Customizability	High levels due to open-source nature, from UI to tracking and reporting.	Limited to predefined settings and features.	Some customization like tagging and categorization.	Minimal customization, focused on app limits and downtime.
Pricing	Free and open-source.	Freemium model, premium version offers additional features.	Free version available, paid license for advanced features.	Free but limited to Apple devices.

Decision Matrix for Use-Cases

Choosing a time-tracking software often depends on specific needs. The following matrix offers suggestions based on different use-cases:

Use-Case	Recommended Software	Reasons
Highly Customizable Solution	ActivityWatch	Open-source nature allows for extensive customization, from UI to tracking metrics.
Privacy-Conscious Users	ActivityWatch	Local data storage by default, with robust privacy controls.
Multi-platform Support	ActivityWatch, ManicTime, RescueTime	All three support multiple platforms (ActivityWatch and ManicTime support Linux too). Apple ScreenTime is limited to Apple devices.
Enterprise Use	RescueTime	Offers team features, data storage on cloud, and centralized reporting.
Blocking Distractions	RescueTime	Provides distraction-blocking features in the premium version. Note that there are other free and open-source tools for blocking distractions, which is why ActivityWatch doesn’t offer this feature.
In-depth Reporting	ActivityWatch, ManicTime, RescueTime	All offer robust reporting capabilities, with ActivityWatch and ManicTime allowing for greater customization.
Apple Ecosystem	Apple ScreenTime	Seamlessly integrates with Apple devices, from iPhones to Macs.
Freelancers on Budget	ActivityWatch	Free and open-source with an extensive set of features for individual tracking.
Offline Activity Tracking	ActivityWatch, ManicTime	Both support offline activity tagging, making it suitable for tracking non-digital tasks.

Conclusion

Time-tracking software options abound, each with its unique set of features, advantages, and limitations. While the final choice will ultimately depend on your specific needs, ActivityWatch emerges as a strong contender in multiple categories.

Strengths of ActivityWatch:

High level of customization thanks to its open-source nature
Strong commitment to user privacy with local data storage
Extensive tracking capabilities, including custom event tracking
Free and open-source, providing a comprehensive feature set without a financial barrier

Competitors’ Strengths:

RescueTime offers distraction-blocking and enterprise features
ManicTime provides a versatile tracking environment, including offline tagging
Apple ScreenTime integrates seamlessly within the Apple ecosystem

In summary, ActivityWatch excels in customizability, privacy, and its free pricing model. It is an ideal choice for users ranging from freelancers to those who prioritize data privacy, all without compromising on tracking capabilities and analytics. If you’re looking for a comprehensive, customizable, and cost-effective time-tracking solution, ActivityWatch may well be the ideal choice for you.

For more alternatives, you can check out alternativeto.net.

ActivityWatch - Open-source time tracker

A lighter faster ActivityWatch with Tauri: now available for testing

Why Tauri

What you get

Developer Experience

Try it today

Who this is for

Help us finish it

Syncing your ActivityWatch devices: what works today, what's next

How it works

What works today

What needs care today

What’s next

Try it, and tell us what breaks

ActivityWatch v0.14.0: faster, steadier, synced, and a new app

Your devices, together

Android caught up

Much faster on large databases

Steadier

A new app, built with Tauri

A refreshed web UI, and more control

Upgrading

Thank you

ActivityWatch Research Edition: a variant for your study

Where it came from

What it does

How much data

Making a variant for your study

Mobile

Get in touch

ActivityWatch has passed 1 million downloads

What’s next

Help us keep going

ActivityWatch for Android 0.14 is stable: sync, alerts, and a lot fewer crashes

Crashes and performance

Sync with your desktop

Alerts, widget, browser tracking

Get it

ActivityWatch vs RescueTime: Why I Built a Free Open-Source Alternative

The Story Behind ActivityWatch

Feature Comparison

Where ActivityWatch Wins

Privacy and Data Ownership

Extensibility

Cross-Platform (Including Linux)

No Paywalls on Your Own Data

Where RescueTime Still Has an Edge

Migrating from RescueTime

What Users Say

Try ActivityWatch

Announcing Winners of Screenshot Competition!

The Winners

🥇 1st place: Erik (@thelifeimliving)

🥈 2nd place: Hernan (@realhyperchango)

🥉 3rd place: @BelKed and @maholli

Prizes

Next Steps

Further Discussion

The Future of Time Tracking: AI, Privacy, and Personalization

Introduction

The Role of AI in Time Tracking

The Power of Data

Long-Term Behavioral Trends

Day Summaries and Note-Taking

In-Context Suggestions

Integrating Brain-Computer Interfaces with Screentime Data: The Next Frontier in Time Tracking

The Synergy of EEG and ActivityWatch

Real-World Applications

Ethical and Privacy Considerations

Future Research and Development

Full-Context Tracking: A Double-Edged Sword

Opportunities for AI Applications

Privacy Concerns

Resource Consumption

Balancing Act: AI, Privacy, and Resources

The Importance of Privacy

The Privacy Imperative

The Resource Trade-Off

The Need for Personalization

Beyond One-Size-Fits-All