EDHOC is a compact and lightweight authenticated key exchange protocol proposed by the IETF, whose design focuses on small message sizes, in order to be suitable for constrained IoT communication technologies. In this tool paper, we overview EDHOC-Fuzzer, a protocol state fuzzer for implementations of EDHOC clients and servers. It employs model learning to generate a state machine model of an EDHOC implementation, capturing its input/output behavior. This model can then be used for model-based testing, for fingerprinting, or can be analyzed for non-conformances, state machine bugs and security vulnerabilities. We overview the architecture and use of EDHOC-Fuzzer, and present some examples of models produced by the tool and our current findings.