# Quick Search turns one bar into apps, files, calendar and a calculator

> Quick Search is an Android search app and launcher written in Kotlin with Jetpack Compose, indexing apps, shortcuts, contacts, files, calendar events, notes, reminders and settings, and adding web search, AI answers, converters, alarms and timers. It runs on Android 7.0 and later, and every permission it asks for is optional.

**teja2495/quick-search** — Universal search app for Android which can search apps, shortcuts, contacts, files, calendar events, notes, settings and the internet.

- Repository: https://github.com/teja2495/quick-search
- Stars: 597 · Forks: 52
- Language: Kotlin
- License: MIT
- Published: 2026-10-22 · Updated: 2026-10-22 · Language: en
- Canonical page: https://hysenlabs.com/projects/teja2495-quick-search

## One search bar over everything on the device

Quick Search is described as a fast, private Android search app and launcher, and the pitch is a single search bar that reaches everything. The list of what it indexes is long enough to be the real specification of the product. Apps, with smart ranking that tolerates typos and matches acronyms such as `yt` mapping to YouTube, plus nicknames, and a long press on an app reaching its shortcuts. App shortcuts, custom deep links and Tasker tasks, shown alongside app results. Contacts, with call and text actions, multi-number support, and WhatsApp, Telegram, Signal and Google Meet actions that can be customised by long press. Files and folders. Calendar events. Notes, with a Quick Note in the widgets panel. Reminders, created in plain language. Notification history, once access is granted. Device and app settings, so you can jump straight to the right settings screen. Screen time, so a search for `Screen Time` shows your daily usage. The point is not any single one of these, it is that a person looking for something does not have to know which of five apps to open.

## Ranking, top matches, aliases and triggers

Four mechanisms sit on top of the raw search and they are what turn a list of results into something usable. The first is smart ranking on apps, which tolerates typos and matches acronyms and nicknames rather than only doing prefix matching. The second is Top Matches, which shows the single best result across all categories, with the categories and their priority configurable, so the answer to a query appears before the list of everything else. The third is aliases and triggers. An alias is a prefix on a query that targets a section, an engine or a tool, and the same mechanism is used to give an engine a short name such as `ggl` for Google. A trigger is set by long pressing any result, which registers a word that opens that result automatically. The fourth is search history, where you can switch between recent queries and recently opened results, which is how a launcher stays useful without retyping. App Suggestions round this out, letting you swipe between New and Updated, Pinned and Most Used, and drag one pinned app onto another to create a folder.

## Web search and AI answers in the same field

The web side is a browser with search built in, and the AI side is a chat box in the same place. There are 30 search engines, including Google, ChatGPT, Gemini, Perplexity, Claude, Kagi, DuckDuckGo, YouTube, Spotify, Reddit and Wikipedia, and you can reorder them, disable them, or assign an alias to any of them. Installed browsers and custom engines can be added too, and web suggestions appear as you type with a configurable count. The AI Search feature gives AI answers with follow-up questions and optional personal context, and it works with Gemini, OpenAI, Claude, Groq, Meta AI and OpenCode, or any OpenAI-compatible provider. A Tavily key can be added to give web results to models that have no built-in search. There is also a Custom AI Tools facility for building your own AI-powered tools, which requires an API key of your own. The practical reading is that the AI feature is a thin layer over whichever provider you point it at, and the privacy posture depends entirely on that provider rather than on Quick Search.

## The built-in tools are what make it a launcher

A search bar that only searches is a search bar, and Quick Search adds a set of small tools that work by typing an expression. The calculator handles math expressions, with calculator buttons on the numeric keyboard. The unit converter covers length, mass, temperature, area, volume, time, speed, data, energy, power, pressure, angle and frequency. The currency converter uses live exchange rates with no API key needed, so `100 USD to EUR` works. Date and time handles natural-language dates, differences, offsets and time arithmetic. Alarms and timers are set with `2:45pm` for an alarm and `15 min` for a timer. Beyond those there is a world clock, weather, a dictionary, a hex and RGB colour visualiser, and Tasker integration. The reminders feature is in the same family, since you create one in plain language such as `Laundry in 2 hrs` and search for it later. This is the pattern worth recognising: every tool is reachable from the keyboard, which is why the app works well as a launcher where a launcher icon would be the alternative.

## It can be your home screen, and it can be a panel over other apps

Quick Search can be set as the launcher, in which case it becomes the home screen and adds app lock, more system app shortcuts, and configurable swipe and double-tap gestures. There is an At a Glance view on the home screen with today's events, upcoming reminders and timers, media controls and low battery warnings. The widgets panel is a swipe-right side panel that holds any Android widget and lets you pin widgets to Home, and the documentation says it works whether or not Quick Search is your launcher. There are home screen widgets too, for search, custom buttons and media controls, with what the README calls extensive customisation. The app also offers several other ways to launch: as the default assistant, through a Quick Settings tile, in overlay mode over other apps, via a system-wide edge swipe, and as a floating button. That last list matters, because it means the tool can reach you without you navigating to it, which is the difference between a search app you open and one that replaces your home screen.

## Permissions, privacy and where the data goes

The privacy section is the one to read before installing, and it is specific rather than decorative. All permissions are optional and only enable the features that need them, and search runs locally on your device. There are no ads and no analytics. API keys are stored encrypted, there are granular permission controls, and data processing is local first. The architecture section backs this up. The language is Kotlin, the UI is Jetpack Compose with Material 3, and the architecture is MVVM with `StateFlow` and unidirectional data flow. Persistence is modular SharedPreferences, encrypted for API keys, and Room for notes and notification history. The widgets are Jetpack Glance. The features list also says there is encrypted storage for API keys and local-first data processing. The caveat is the AI Search feature, which by definition sends your query to the provider you configured, and the web search engines, which send it to a search engine. So the local-first claim holds for device search and falls away for the two features that need a network.

## Installing it: three channels, one requirement

The requirement is Android 7.0, API level 24, or higher. There are three ways to get the app. Google Play gives you automatic updates. F-Droid gives you repository-based updates. And the Releases page on GitHub carries the latest APK. If you want to build it yourself, the repository is a Gradle project with a wrapper, and there are two release variants.

```bash
git clone https://github.com/teja2495/quick-search.git
cd quick-search

# Google Play / GitHub releases (includes Play in-app review & updates)
./gradlew assembleStandardRelease

# F-Droid build (no Google Play libraries bundled)
./gradlew assembleFdroidRelease
```

The two variants are the interesting part, and the difference is stated in the comment itself. The standard release includes the Google Play in-app review and updates machinery, which means it carries the Play libraries. The F-Droid build has no Google Play libraries bundled, which is what a reproducible F-Droid build requires, and the publishing process is documented separately in `docs/FDROID.md`. The repository also carries a `fastlane/` directory with the screenshot metadata, a `benchmark/` module, a `DESIGN_SYSTEM.md`, a `PRIVACY_POLICY.md` and an `AGENTS.md`. The release history is steady, with 4.5 on 2026-09-18, 4.6 on 2026-09-22 and 4.7 on 2026-09-29, and the last push was on 2026-10-18.

## Conclusion

Adopt Quick Search if you want a single search field on Android that covers apps, files, calendar, notes and settings, and you are willing to grant the optional permissions that open up the deeper categories. Skip it if you want a plain launcher with no data access, since the app's value comes from indexing what is on the device, and skip it if you do not want an app that can query an AI provider from a search bar. Before you install, read the privacy policy and the permissions list, then build the F-Droid variant if you want a build with no Google Play libraries bundled.

## FAQ

### What can Quick Search on Android actually search?

Apps with typo tolerance and acronym matching, app shortcuts and Tasker tasks, contacts, device files, calendar events, notes, reminders, notification history, device and app settings, and screen time, with a Top Matches view for the best result across all categories.

### Which AI providers does the Quick Search AI feature support?

It works with Gemini, OpenAI, Claude, Groq, Meta AI and OpenCode, or any OpenAI-compatible provider. A Tavily key can be added to supply web results for models without built-in search, and custom AI tools can be built with your own API key.

### Does Quick Search need any permissions to work?

All permissions are optional and only enable the features that need them, and search runs locally on the device. The privacy section also states there are no ads or analytics, encrypted storage for API keys and local-first data processing.

### How do I build Quick Search from source for F-Droid?

Clone the repository, change into the directory and run `./gradlew assembleFdroidRelease`, which produces a build with no Google Play libraries bundled. The standard release is built with `./gradlew assembleStandardRelease` and includes the Play in-app review and updates.

## Sources

- [Issues](https://github.com/teja2495/quick-search/issues)
- [License: MIT](https://github.com/teja2495/quick-search/blob/main/LICENSE)
- [README](https://github.com/teja2495/quick-search/blob/main/README.md)
- [Releases](https://github.com/teja2495/quick-search/releases)
- [teja2495/quick-search on GitHub](https://github.com/teja2495/quick-search)

---

Hysen Labs editorial analysis, written from the project's own repository and release notes. Cite the canonical page: https://hysenlabs.com/projects/teja2495-quick-search