PyMC is a Python library and open-source project that helps users build and analyze Bayesian models through a straightforward, Python-based interface. Designed to make Bayesian modeling more accessible, it lets practitioners focus on their questions while providing tools to specify models and draw conclusions from data. PyMC supports modern inference methods, including Markov chain Monte Carlo (MCMC), with the NUTS sampler, and variational inference through ADVI.

The library includes probability distributions, Gaussian processes, approximate Bayesian computation (ABC), and sequential Monte Carlo (SMC). Its PyTensor computational backend can compile models using C, Numba, or JAX, support GPU execution, and apply graph optimizations. PyMC also integrates with ArviZ for visualization and diagnostics, and with Bambi for high-level mixed-effects models.

The project website provides installation guidance, beginner and API guides, example notebooks, and community resources such as a discussion forum and meetups. PyMC is maintained as a community-focused project, with opportunities to contribute and connect through its online channels.