Features:

Data (such as volume and pan settings) is accessed by the user in plain text files.
Can be run without a desktop, a windows manager, or a soundcard.
It is virtually unbreakable by other code/dependencies since it uses only the Alsa library (on Linux) and built-in C libraries.
Integer math summing for precision.
Direct signal paths and minimal processing whenever possible.
Built in 64-bit precision equalizer, compressors, and reverb.
Fast offline rendering with preview mode for hearing sections only.
Panning works on principle of time delay or user settable pan law.
Volume automation and ramps.
127 tracks and 8 stereo submixes.
Works on 32-bit or 64-bit operating systems.
Run simultaneous mixes or submixes on different cores.
Installs simply by copying the very small binary to a working project or applications directory.
User choosable audio file time placement (time stamping).
Beat or time based automation entry with adjustable tempo and meters (tempo mapping).

Limitations:

Audio playback is limited to the Linux version.
Reads and writes 16 and 24-bit .wav files only.
No editing or recording capability (see Recorder7 on main page).
Does not integrate with other software plug-ins.
Automation is limited to track volumes.
Works offline only: workflow is adjust, render, then listen.
Effect ordering is fixed.

Changes in v2.10(r16)

• Major engine work: disk read/write size is now independent of ALSA buffer size for added stability with low ALSA buffer.
• Added user ability to tweak timeline placement on overdubs.
• Fixed record-stop timing that sometimes ended slightly early.
• Fixed bug that prevented a new take from being auditioned more than once.

Changes in v1.0953

• Improved clock display during playback
• Increased playback buffer
• Added some verbosity (ie. minimal text) function during playback