Logo ot the FQLite
          Project

Forensic SQLite Data Recovery Tool 

Author: Dirk Pawlaszczyk


Introduction


FQLite is a tool to find and restore deleted records in SQlite databases. It therefore examines the database for entries marked as deleted. Those entries can be recovered and displayed. It is written with the Java programming language. The program can operates with a simple graphical user interface (GUI mode). The program is able to search a SQLite database file for regular as well as deleted records.

Technical Background


SQLite is a software library that implements a simple SQL database engine. It is trendy and used in different software products. Mozilla Firefox and Google Chrome, for example, use SQLite version 3 databases for user data such as history, cookies, and downloads. Sometimes single records are deleted from the database, or even a complete table has been dropped. But this does not mean that this data is physically removed. In most cases, such artifacts can still be found, even if the beginning of the data records has been partially overwritten. Beyond this, a database file might contain one or more pages that are not in active use. Such unused pages can arise when information is deleted from the database. They are stored on the free list and were reused, whenever new pages are required. Such artifacts are often of high forensic value. FQLite can make this data visible again by extracting the desired information directly from the binary data. The tool works without an interface on the raw data image.


The User Interface
            (Main-Screen)
Figure: The FQLite user interface

Features



FQLite allows you to:
Some features:


user interface
Figure: The SQL Analyzer

Text-to-SQL
              Assistant
Figure: Text-to-SQLAssistant


Benchmark


Tool comparison 2026: eight recovery tools

In a new, fully reproducible benchmark, FQLite and seven other open-source recovery tools (sqlite4n6, DC3 SQLite Dissect, bring2lite, SQLite GUI Analyzer, NFI xsqlite, sqlite_extract and SQLite's built-in .recover) were run on the complete SQLite Forensic Corpus [1] – 94 databases including all nine anti-forensic categories – and on the false-positive scenarios of Lee et al. (2025). All tools were scored by the same code against the ground truth of the corpus [3].

Bar chart: share of deleted records recovered and of live rows
              recovered from manipulated databases, for eight SQLite recovery tools
Figure: Deleted records recovered (recall, 234 recoverable records in categories 0A–0E) and live rows recovered from manipulated databases (5,873 rows in categories 11–19)

FQLite recovered 232 of 234 deleted records without a single false positive, recovered 98.2 % of the live data in manipulated files and completed every run without crash, time-out or memory exhaustion. The benchmark also revealed defects in FQLite itself; they are fixed in release 5.11.

Tool Recall
(deleted)
Precision
(false positives)
Live data
(manipulated)
Failed runs
(of 94)
FQLite0.9911.000 (0)98.2 %0
SQLite GUI Analyzer0.9910.996 (1)95.5 %0
sqlite4n60.8251.000 (0)n/a0
bring2lite0.5600.963 (5)89.3 %34
DC3 SQLite Dissect0.3970.123 (665)7.8 %62
NFI xsqlite0.2950.084 (750)0.0 %64
sqlite_extract0.0510.375 (20)94.8 %0
SQLite .recover0.0511.000 (0)94.5 %0

Table: Results of the 2026 tool comparison (D. Pawlaszczyk 2026 [3], data and harness [4]). The paper is currently under review at ScienceOpen.

Earlier results (2021)

Below are some test results of programs achieved with the forensic corpus [1], a benchmark specifically designed for SQLite databases. The table shows how well the individual tools performed in restoring the 278 deleted records of the corpus.

Undark SQLite
Deleted
Record
Parser
Stellar
Phoenix
Repair
for SQLite
SysTools
SQLite
Database
Recovery
Sanderson
Forensic
Browser
for SQLite
95/278
34,8%
139/278
50%
0/278
0%
0/278
0%
33/278
11,9%
SQLabs
SQLite
Doctor
Sqlite
Forensic
Explorer
Autopsy
SQLite
Deleted
Records
Plugin
bring2lite FQLite
0/278
0%
73/278
26,3
0/278
0%
147/278
52,9%
278/278
100%

 Table: Recovery rates of FQLite compared to
9 other tools tested with the Forensic Corpus
(vgl. D. Pawlaszczyk and C. Hummert 2021 [2],  S. Nemetz and S. Schmitt et al 2018 [1])

Download latest version



November 2026

GitHub Downloads (all assets, specific tag)  GitHub Downloads (all assets, specific tag)  GitHub Downloads
            (all assets, specific tag) 

The latest stable version 5.11 is available for download here:

click
                here to start download the Linux version for X64
                architecture click
                here to start download the Windows version for X64
                architecture click
                here to start download the MAC OS version for X64
                architectureclick here to start download the MAC OS version for
                M1/M2 Macs


Changes in release 5.11:
  • many improvements to the robustness of the import process
  • several anti-forensic manipulations are now handled (page pointers, freelist trunks, cell metadata, overflow chains)
  • bug fix for WAL import (only one block was read)
  • improved overflow handling; incomplete recovered records are flagged
  • reworked B-tree analysis
  • icons are now mainly SVG-based
Changes in release 5.1:
  • NodeQL integration: visual, block-based SQL query builder with inline results
  • shared light/dark theme
  • bug fixes for WAL archive and rollback journal analysis
Changes in releases 5.0, 4.22, 4.21 and 4.01:
  • support for CDR analysis (Lighthouse plugin) (5.0)
  • support for SQLCipher databases; customizable timestamp format (4.22)
  • CASE/UCO export (4.21)
  • bug fixes for WAL archive and rollback journal analysis, LLM path handling and master-table parsing (4.01–5.0)
Changes in release 4.0:
  • the schema view was completely reworked
  • HTML Report functionality was added
  • introducing a new RAG-pipeline support AI-based a Text-to-SQL functionality for SQL query generation
  • SQL-Analyzer now works with SQLite syntax, Support for other dialects has been canceled
  • smaller Bug fixes
  • solved some issues around IllegalStateException: "Not on FX application thread"

Changes in release 3.3:

  • an export into a SQLite database is supported
  • reworked user interface (introduced new icons, added points to context menus)
  • source code cleaning and improvments - better readability, comments were added
  • login level can now be adjusted by the user within the settings dialogue
  • smaller bug fixes

Changes in release 3.2:

  • smaller bug fixes in detection algorithm
  • added a new ROWID column
  • pll and hl columns are merged now into one column
  • UI layout improvements

Changes in release 3.1:

  • SQLAnalyzer beta has been added
  • HexViewer -> bug fixes (blank lines when klick on offset)
  • Improvements during db import (FXObservable)
  • Licence info was updated (changed to Apache 2.0 Licence)

Changes in release 3.0:

  • We now support ".exe" installation files for Windows as welle as ".deb installation" package for Linux
  • Updated user handbook
  • smaller bug fixes

Changes in release 2.7:

  • Export functionality has been completely reworked (new dialog window)
  • Font settings as well as export settings are now persistent
  • Fixed bug in multi-row selection


Changes in release 2.65:

  • Added support for FTS3/FTS4 virtual tables
  • Massive speed improvements for larger files (including WAL)
  • Solved some issues with software locks in GUI mode
  • Bug fixes in WAL overflow handling
  • Encrypted databases (e.g. Signal.db) are now automatically detected and import stops
  • smaller bug fixes for BLOB preview

Changes in release 2.6:

  • Added support for databases with a size > 2,1 GB
  • Hex-Viewer is now only displayed if needed (not allways in Background)
  • There is a new ViewLog option in the 'Info'-Menu - you can now check the log-messages
  • Columns for header fields, schema-info  and pages-tab are new sortable
  • Filter function on table view has been updated
  • Links in the Schema-Detail-View now work as expected
  • Fonts-Dialog now offers Support to go back to Default Font
  • Smaller bug fixes 

Changes in release 2.5:

  • Improved support for decoding BLOB-columns (BSON, MessagePack, AVRO, Thrift...)
  • Smaller bug fixes

Changes in release 2.41:
  • Improved detection of deleted files
  • New pages tab for preview page types
  • Improved Filter-function for tables (you can now filter for a particular column)
  • Several smaler bug fixes

Changes in release 2.3:

  • Support for pdf-BLOBs with internal PDF-Viewer
  • copy Tooltip content to clipboard
  • view .tiff and .heic BLOBs with OS default viewer fort those file types
  • clicking on the <help> - function will open an PDF-Version from the user manual (before, the online help was displayed)

Changes in release 2.2:
  • Preview of bplist (Apple plist-format), BASE64 decoded string as well as Google ProtoBuffer (experimental) is now supported
  • Improved timestamp support for Integer and Real fields (in tooltip)
  • Easy BLOB export to clipboard
  • Some smaller bug fixes for data schema analyzes (schema table definitions without a sql-type are now correctly mapped to BLOB) 


In the latest version, the FQLite is bundled with a Java Runtine Environment (JRE) and all required libraries.

Important note: With version 2.0 the support for the command line mode was cancelled.

User Guide


The latest version  of the official userguide can be obtained here

SQLite Forencis with FQLite - The Official
                Userguide


Installation Instructions



macOS
Installation via .dmg File


1. Download the latest version of FQLite in .dmg format from the Release page.
2. Open the .dmg file and drag the application into the "Applications" folder.
3. You can now launch FQLite from the Applications folder.

Important node: If you try to open an app by an unknown developer and you see a warning dialog on your Mac. A dialog is displayed saying that the app is damaged. In fact, the app is simply not signed with a developer certificate. For this reason, Gatekeeper refuses to execute. The first method will allow a single program to run, without having to disable Gatekeeper. Open a terminal and run the following command:
sudo  xattr -dr com.apple.quarantine /Applications/fqlite.app 

The app should then start without any further complaints.

Windows
Installation via .exe File


1. Download the latest version of FQLite in .exe format from the Release page.
2. Run the .exe file and follow the installation instructions.
3. After installation, FQLite can be opened from the Start menu.

Linux
Installation via .deb File

1. Download the latest version of FQLite in .deb format from the Release page.
2. Open a terminal and navigate to the directory where the .deb file is saved.
3. Install FQLite with the following command:

sudo apt install ./fqlite.deb

4. After installation, FQLite can be launched from the application menu.


Video tutorial and other resources



You can watch here a small introduction video, which explains the essential functions of FQLite right here (in German) click here.

To download an example database with 9 different binary format samples click here.

To view a short video showing the decoding of binary data with FQlite click here.

Get the Source Code



You check out the source code you can download the github repository from this link:

https://github.com/pawlaszczyk/fqlite

GitHub Logo

References


[1] S. Nemetz, S. Schmitt, F. Freiling: A standardized corpus for SQLite database forensics. In: Digital Investigation, vol. 24, Supplement, 2018, pages 121-130, (2018).    
https://doi.org/10.1016/j.diin.2018.01.015.

[2] D. Pawlaszczyk, C. Hummert: (2021). Making the Invisible Visible – Techniques for Recovering Deleted SQLite Data Records. International Journal of Cyber Forensics and Advanced Threat Investigations, 0, 1-1. Retrieved from https://conceptechint.net/index.php/CFATI/article/view/17

[3] D. Pawlaszczyk: Recovering Deleted and Manipulated Records from SQLite Databases: A Reproducible Benchmark of SQLite Forensic Tools. Preprint, ScienceOpen, 2026 (under review).

[4] D. Pawlaszczyk: SQLite Recovery Benchmark: harness, test data and results for eight forensic tools. Zenodo, 2026. https://doi.org/10.5281/zenodo.23928446

[5] D. Pawlaszczyk, R. Bodach, P. Engler, J. Kolouch, M. Spranger, C. Hummert, D. Labudde: AI-based automated SQL query generation for SQLite databases in mobile forensics. Forensic Science International: Digital Investigation, vol. 57, 2026, 302100. https://doi.org/10.1016/j.fsidi.2026.302100

Also check out my publication on Springer Mobile Forensics – The File Format Handbook. This open access book summarizes knowledge about several file systems and mobile file formats commonly used in mobile devices. It also includes a chapter on Forensic Analysis of SQLite Databases.

File Format
                  Handbook - Cover


Contact



Address:
Dirk Pawlaszczyk
Technikumplatz17
09846 Mittweida
Germany

Email:
[email protected]

My university web page:

https://www.cb.hs-mittweida.de/professoren-innen/forensik/prof-pawlaszczyk/

Contact me on Researchgate:

https://www.researchgate.net/profile/Dirk_Pawlaszczyk

 This web page was tested and can be viewed with the Netscape Navigator 9 and Lynx web browser.  No tracking. No JavaScript.

© 2019-2026 Dirk Pawlaszczyk 
Impressum