The use of Multi-Rotor Aerial Vehicles (MRAVs) in tasks that require physical interaction has been an active research field in the last decade which resulted in an increasing interest in Aerial Manipulators (AMs). This raises many challenges in the modeling, control, perception, and planning of these robots. However, designing and realizing an AM testbed is a complicated multi-disciplinary task, and there is a lack of standardization in the relatively new field of AMs. For this purpose, we introduce Sarax, an open-source hardware and software framework tailored for AMs research and innovation. The software of Sarax is built on top of open-source projects such as the Robot Operating System (ROS) and PX4 Autopilot, while the hardware is designed to be customizable, modular, and easily scalable through parameterized models. We verified and validated the proposed framework through indoor and outdoor experiments. We aim to open the door to accelerate AMs research and innovation, allow researchers and developers to focus on their core contributions, and take AMs technology to a higher readiness level.