The low-level data processing of the Cherenkov Telescope Array (CTA) and indeed all other exist- ing Cherenkov Telescopes can be broken into four general steps: 1) the processing of air-shower event image time-series, 2) the stereo reconstruction of the incident air showers, and 3) the dis- crimination of gamma-ray induced showers from those from cosmic rays 4) the determination of the overall system response. The final output for science users is a list of reconstructed gamma-ray-like events and their associated parameters, along with a set of instrumental response functions needed for doing astrophysics. We present a python-based framework, ctapipe, for writing the algorithms required for these processing steps along with a reference prototype pipeline. The code is written with a focus on simplicity and usability by developers with a diverse range of skill sets, and leverages existing code from the science community (AstroPy, SciPy/NumPy, SciKit- Learn, etc). This concept is intended to be a prototype for the final CTA low-level data processing pipeline, allowing physicists to quickly explore low-level Cherenkov telescope data and develop new algorithms. Thanks to the framework modularity, computer engineers and data scientists will be able to simultaneously optimize the algorithms and parallelize them using modern computing and big-data architectures to support the high data volumes of CTA.