Interpreter
Interpreters are programs that take in source code and executes it directly without compiling it to machine code. Instead, a common way interpreters are implemented is to parse source code, compile it to an intermediate representation (often called bytecode), then hands that off to a virtual machine to execute.
Interpreters may or may not use a JIT compiler.
# Interpretation vs Compilation
Interpretation and compilation are not mutually exclusive and are not simply alternatives of each other. Likewise, interpreted languages and compiled languages don’t necessarily mean that they must be only put through an interpreter or compiler respectively.
Interpretation is just “given this source code, do what it says”, which is basically just executing the source code directly (from the programmer’s perspective). Compilation is a mapping from language A to language B, like converting C++ code into x86 machine code.
When we say a language is a compiled language or interpreted language, it really just means that the canonical or official compilation model intended by the language designers is to have the language be fed through a compiler or interpreter, respectively. Whether or not a language is compiled/interpreted is not a property that’s tied to the programming language, but rather how it’s implemented. A compiled language like C++ does actually have interpreters you can use for it, and an interpreted language like Python does have compilers.
You can have both compilers and interpreters for the same programming language.