Linux file search, instant results as you type
Alternatives To Angrysearch
Project NameStarsDownloadsRepos Using ThisPackages Using ThisMost Recent CommitTotal ReleasesLatest ReleaseOpen IssuesLicenseLanguage
2 hours ago138mitRust
A lightning-fast search engine that fits effortlessly into your apps, websites, and workflow.
20 days ago24July 10, 202266mpl-2.0Rust
🦔 Fast, lightweight & schema-less search backend. An alternative to Elasticsearch that runs on a few MBs of RAM.
13 hours ago280gpl-3.0C++
Open Source alternative to Algolia and an Easier-to-Use alternative to ElasticSearch ⚡ 🔍 ✨ Fast, typo tolerant, in-memory fuzzy Search Engine for building delightful search experiences
a month ago33April 21, 2021130mitGo
Real-time Geospatial and Geofencing
Awesome Elasticsearch4,408
6 months ago2unlicense
A curated list of the most important and useful resources about elasticsearch: articles, videos, blogs, tips and tricks, use cases. All about Elasticsearch!
Search Replace Db3,6222462 years ago6May 12, 20202gpl-3.0PHP
This script was made to aid the process of migrating PHP and MySQL based websites. Works with most common CMSes.
a day ago221gpl-2.0C++
Easy to use open source fast database for search | Good alternative to Elasticsearch now | Drop-in replacement for E in the ELK soon
Js Search1,438164883 years ago21March 13, 20207mitJavaScript
JS Search is an efficient, client-side search library for JavaScript and JSON objects
Walrus1,05226174 months ago39June 30, 2022mitPython
Lightweight Python utilities for working with Redis
a year ago37gpl-2.0Python
Linux file search, instant results as you type
Alternatives To Angrysearch
Select To Compare

Alternative Project Comparisons


Linux file search, instant results as you type

Attempt at making Linux version of Everything Search Engine because no one else bothered.
Everyone seems to be damn content with searches that are slow, populating results as they go; or are cli based, making it difficult to comfortably make use of the results; or are heavily integrated with a file manager, often limiting search to just home; or are trying to be everything with full-text file's content search.

A similar project worth attention - FSearch

demonstration gif

Done in python 3 using PyQt5 for GUI

Lite mode vs Full mode

angrysearch can be set to two different modes in its config, default being lite

  • lite mode shows only name and path
  • full mode shows also size and date of the last modification, the drawback is that crawling through drives takes roughly twice as long since every file and directory gets additional stats calls

in ~/.config/angrysearch/angrysearch.conf you control the mode with angrysearch_lite being set to true or false

lite version png

Search modes

there are 3 search modes, default being fast

  • fast mode - enabled when the checkbox next to the input field is checked
    extremely fast, but no substrings, meaning it would not find "Pirates" or "Whiplash", but it would "Pirates" or "The-Fifth"
  • slow mode - enabled when the checkbox is unchecked, slightly slower but can find substrings, also very litteral with non typical characters
  • regex mode - activated by the F8 key, indicated by orange color background
    slowest search, used for very precise searches using regular expressions, set to case insensitive,
    unlike the previous search modes not entire path is searched, only the filenames/directory names

regex example:

regex in action gif

What you should know:

  • the database is in ~/.cache/angrysearch/angry_database.db
    the config file is in ~/.config/angrysearch/angrysearch.conf
  • it can take ~2 min to index ~1 mil files, depending on hdd/ssd and filesystem - ntfs on linux being much slower. The database might be ~200MB
  • it is not recommended to run as root, there's no reason for it and you might crawl where you would rather not, like Btrfs users going in to snapshots
  • xdg-open is used to open the files based on their mimetype, default applications can be set in ~/.local/share/applications/mimeapps.list or ~/.config/mimeapps.list


xubuntu installation demonstration

Manual installation is easy as there's no compilation with python, process consists of having dependencies, copying files somewhere and setting execution permissions

dependencies - python3-pyqt5, xdg-utils

     you need PyQt5 for python3, for example ubuntu based distros: sudo apt install python3-pyqt5
     most distros have xdg-utils out of the box

Now that you have the dependencies, download the latest release of angrysearch and unpack it somewhere. Along the files there's one called, it will copy files where they belong and sets correct permissions.

  • open terminal in the directory with the release files

  • set as executable and run it

      chmod +x
      sudo ./
  • DONE, if you want to see more detailed instruction, here is older version of this readme


  • python3-gobject - desktop notifications for automatic update, most DEs have it
  • xdotool - needed if using Thunar or PCmanFM and making use of the config option fm_path_doubleclick_selects

Automatic update in the background

notifications png

Among the files there's
When this file is run there's no interface, it just crawls through drives and updates the database

Using crontab you can set this file to be executed periodically at chosen intervals, keeping angrysearch up to date with the changes on your system

  • crontab -l - list cronjobs
  • crontab -e - open text editor so you can enter new cronjob

this cronjob will execute the update every 6 hours

0 */6 * * * /usr/share/angrysearch/

crontab does not try to catch up on a job if the PC has been off during scheduled time

notifications setting turns on/off desktop notifications informing about automatic update finishing
conditional_mounts_for_autoupdate can prevent autoupdate from running if set mount points are not present

Desktop notifications from cronjob not always work, so on your distro you might be without them

How it works & additional details:

look in to the database

  • On update angrysearch crawls through your file system and creates its database.
    The database has a column containing full path to every file and directory found, another column indicates if the path is to a file or a directory. If full mode is enabled then there are also columns for the last modification date and for the size of files in bytes.

  • When typing in to the search input the path column is searched for the occurrences of typed text and the rows containing them are shown.
    This is unlike other searches which usually look only through names of files/directories not their entire paths. This means that writing books will show all the items with the term "books" somewhere on their path instead of just in the name.
    On typical slow searches this would be too broad of a search with too many results, but the instantaneous nature of angrysearch allows to continue typing until the search is narrow enough.

  • The database uses FTS extension of sqlite for indexing to dramatically improve search speed and get the instantaneous feel - results as you type - fast mode
    Drawback of this indexing is inability to do substring searches, but the checkbox in the top right corner can change this. If it's unchecked it will not use FTS tables and just do regular slower database search query - slow mode

  • In the fast mode quotation marks can be used to make exact searches: 'torrent' would not include "torrents" in the results.

  • file alone is all that is needed for full functionality. But no special icons or dark theme.

  • Hovering mouse over the update button will show how old is the database.

  • double-click on the items in search results:

    • Name - the first column, opens the file in the application associated with its mimetype using xdg-open
    • Path - the second column, opens the item's location in the file manager
  • Results can be sorted by clicking on column's headers, only the presented results will be sorted, meaning that by default max 500 items. To return to the default sort, sort by path column.

  • Hotkeys

    • F6 ctrl+L alt+D - focus search input
    • Enter in search input- jump to results
    • Enter in search results - open selected item in associated application
    • shift+Enter - open items location
    • Tab - cycle through UI elements
    • shift-Tab - cycle backward through UI elements
    • arrow up arrow down - navigate through search results
    • Esc ctrl+Q - exit the application
  • FTS5 is the new version of the indexing extension of sqlite, most distros don't have it yet and are on FTS4. The systems that do have it get two additional benefits in the fast mode

    • can exclude from search results by using the minus sign: search wav -home would show all paths containing the word wav except the ones also containing home
    • ignorance of diacritic, search for oko would also show results like ôko ókö Okǒ

    To check if FTS5 is available on your system - in update dialog window, hover mouse over the text • creating new database


  • config file location: ~/.config/angrysearch/angrysearch.conf
    You can delete the config file whenever you wish, on the next run/close a new one will be created with the default values.

config file screenshot

  • angrysearch_lite By default set to true. In the lite mode theres only file name and path, no file size and no last modification date. Less informations but faster crawling through the drives
  • close_on_execute By default set to false. Closes angrysearch after opening a file or a path in a file manager
  • conditional_mounts_for_autoupdate By default empty. Purpose is to hold mount points that should be present when the database is being updated. If a mount is missing, automatic update through crontab will not run, but use system notification dialog to inform that paths set in this settings are not mounted. This prevents overwriting the database when not all drives are present. Values are system mount points, space separated.
  • darktheme By default set to false. If set true dark theme is used for the applications interface, as defined in the qdarkstylesheet.qss, also contains icons for dark theme
  • directories_excluded By default empty. Which directories to be ignored. Just name of the directory will ignore every directory of that name, full path like /var/cache/pacman/pkg/ ignores exactly that single folder, or parent/target for more easily targeting specific folder pacman/pkg. Can be set through program's interface, in the update window. Directory /proc is hard coded to be ignored
  • fast_search_but_no_substring By default set to true. It holds the last set value of the checkbox affecting the speed of search and substrings, see FTS4 in the section above
  • file_manager By default empty. Whatever application/script is put there, it receives the path when the path column is double-clicked. If left empty angrysearch will try to autodetect default file manager using xdg-utils. If one of the following file managers are set/detected: ['dolphin', 'nemo', 'nautilus', 'doublecmd'], the behavior will change slightly, sending to those file managers full path to the file, highlighting the target file when opened in a file manager.
  • fm_path_doubleclick_selects By default set to false. Needs xdotool package, and preferably manually set file manager in config. When set to true, Thunar, PCmanFM and SpaceFM file managers will be able to open containing directory with the file selected
  • icon_theme By default set to adwaita. Which icon theme to use, can be set from program's interface in the update window. There are 6 icon types - folder, file, audio, image, video, text. Did not yet figure out how to get theme of the distro and reliably icon from file's mimetype, so packing icons with the angrysearch is the way
  • notifications By default set to true. Automatic periodic updates that are run on background using crontab will use desktop notification system to inform when crawling is done or if it was aborted because of missing mount points
  • number_of_results By default set to 500. Limit set for searches in the database. Lower number means search results come faster
  • regex_mode By default set to false. Enables regex search mode. F8 key toggles between true/false when running the application
  • row_height By default set to 0 which means default system height. Sets height of the rows in pixels
  • typing_delay By default set to false. If enabled, it introduces 0.2 second delay between the action of typing and searching the database. This will prevent unnecessary database queries when user is typing fast as there is waiting to finish typing. This can improve performance on slower machines, but on modern ones it might negatively affect the feel of instant responsiveness
  • [Last_Run] The applications properties from the last time at the moment when it was closed - window size, position, state

dark theme screenshot

Popular Search Projects
Popular Database Projects
Popular Computer Science Categories

Get A Weekly Email With Trending Projects For These Categories
No Spam. Unsubscribe easily at any time.
Desktop Notifications