Frida was born after @oleavr and @hsorbo had been casually brainstorming about how they wish they could turn tedious manual reverse-engineering into something much more fun, productive, and interactive.
Having built oSpy and other custom tools to scratch reverse-engineering itches, @oleavr started piecing together frida-gum, a generic cross-platform code-instrumentation library for C. At the time it was limited to hooking functions and providing some tools to help developers write unit-tests for memory leaks and profiling on an extremely granular level. Later it was further improved and used to create Frida. The component frida-core would take care of all the nitty gritty details of injecting shared libraries into arbitrary processes, and maintaining a live two-way channel with the injected code running inside those processes. Inside that payload, frida-gum would take care of hooking functions and providing a scripting runtime using the excellent QuickJS engine.
Later, in their not-so-ample spare time, @oleavr and @karltk did some recreational pair-programming-hackathons that resulted in huge improvements to frida-gum’s code tracing engine, the so-called Stalker. There were also Python bindings created. They started realizing that it was about time that people out there knew about the project, so further hackathons were devoted to piecing together a website and some much needed documentation.
Today, Frida should be a very helpful toolbox for anyone interested in dynamic instrumentation and/or reverse-engineering. There are now language bindings for Node.js, Python, Swift, .NET, Qt/Qml, and it is also possible to use Frida from C.