This article provides an overview of witnesses and the role they play in stateless Ethereum. It examines the current state of witness-related initiatives, and their overall projected use in the Ethereum protocol. It aims to provide a comprehensive review of the topic without straying into client-specific implementation details. As the protocol has not migrated to Verkle trees yet for storage, this document stops short of discussing how Verkle witnesses would look and operate. Rather, it concentrates on the Merkle tree versions that are currently available, the interplay with the concept of access lists, and some potential improvements and applications that do not require Verkle trees to be finished. What is a witness? A witness proof? The term "witness" in mathematics or cryptography refers to a piece of information that enables the verification of some statement. As a basic example, consider a value y, a one way hashing function HASH(), and a witness x. We can easily check the truth of the statement y = HASH(x) by evaluating the right hand side and comparing to the left. Here, x is the witness, and HASH() acts as a sort of verification function. In the context of Ethereum and particularly statelessness, the witness has a more specific definition. It commonly refers to sections of a Merkle (or other) tree that will allow us to answer the question of whether a certain value exists in this tree. In this scenario, the witness (set of hashes from the state tree) acts as "evidence" of the values. The verifier is an algorithm or other computation that performs some operations on the witness and concludes whether or not the value it is concerned with can exist in this tree. What is a witness in the context of Ethereum? We know that we can think of Ethereum as a state machine. Each client stores a copy of the global (world) state as a Merkle Tree. When a block is processed for execution and validation, account information in the state tree is written and read according to the transactions in the block.