Squey is designed from the ground up to take advantage of GPUs and CPUs to perform interactive explorations of massive amounts of data.
It gives users an exhaustive yet intuitive multi-view representation of columnar data and can ingest from:
- Structured text files (CSV, logs, ...)
- Apache Parquet files
- Pcap files
- SQL databases
- Elasticsearch databases
Squey strives to deliver value through its V.I.SU approach:
- Visualize: Leverage various visual representations of raw data in combination with statistics
- Investigate: Use filters to build an accurate understanding of millions of rows while switching instantly between capturing the big picture and focusing on the details
- Spot the Unknown: As a structured understanding of the data emerges, identify unknowns and anomalies
Squey can be used for many different purposes, such as:
- BI and Big Data: Bootstrap initial understanding of complex datasets and deep dive where necessary to design accurate data processing
- Cybersecurity: Detect weak signals such as attacks and data leaks
- IT troubleshooting: Resolve network issues and improve application performance
- Machine Learning: Design training dataset to fulfill targeted improvements of Machine Learning models
Give yourself a chance to see your data and have fun exploring!
Screenshots
Observed history
| App | Change | Version | When ↑ |
|---|---|---|---|
| updated | 5.2.0 | 8 Oct 2026 |
Upstream releases
Release notes supplied by the developer. These dates are not Flathub package publication times.
5.2.08 Oct 2026
This major release adds the following features:
- scatter thumbnails gallery
- SQL console with autocompletion
- undo/redo with breadcrumbs
- selection stretch over the whole axes of the parallel view
- lines density opacity in the parallel view
- antialiased lines in the parallel and zoomed view
- [linux] support rendering on AMD and Intel GPUs
- draw each group of a split series in its own colour
- fix searching an IP address selecting the rows without address
- fix dates read in the time zone of a previous date
- fix PCAP fields kept in all their occurrences coming out empty
- fix the values printed at the ends of the axes being swapped
- fix opening CSV formats and investigations saved by 5.1.1 to 5.1.3
- fix a crash when importing a file with a guessed format
- fix crashes when closing views and removing sources
- fix a crash when closing a file dialog
- fix opening investigations whose format holds non-ASCII characters
- fix a hang when switching the axis of a zoomed view
- fix a hang when removing a column
- rendering up to 10 times faster on the processor without GPU
- compile the rendering kernels behind the splash screen
- fix opening compressed files with "open with"
- open the files a second launch hands over to the running Squey
- fix the low memory warning raised while importing
- speed up computations on a selection up to 5 times
- speed up the building of zones and the series view
- speed up reading and writing dates to the microsecond
- read more dates to the microsecond
- prefer a dedicated GPU over the processor
- report the error tshark gives when it refuses a capture
- [linux] declare the file types Squey opens, for "open with"
- [linux] render through OpenCL backend instead of QPainter
- [linux] fix invalid values lost on file systems that cannot allocate
- [macos] fix a crash when importing columns with invalid values
- [windows] keep the compiled rendering kernels across reboots
- [windows] ship 7z.dll so that 7-Zip no longer needs installing
- [windows] fix selecting time ranges longer than half an hour
- [windows] read the UTC offsets of dates
- [windows] fix a handle leaked by every file mapping
- [windows] divide by 8 the memory taken by selections
- update freedesktop runtime from 25.08 to 26.08 and all libraries to latest versions