What Stan does

Stan is an open-source probabilistic programming language for Bayesian statistical inference, founded in 2011 as a research project at Columbia University and now operated as a global community project under NumFOCUS, a U.S. 501(c)(3) nonprofit. Its core technology is a domain-specific Bayesian modeling language compiled by stanc3 (an OCaml compiler) to C++ executables, with a C++ mathematical library (Stan Math) that provides automatic differentiation powering Hamiltonian Monte Carlo (HMC) and the No-U-Turn Sampler (NUTS) for posterior sampling. The platform supports a wide range of modeling applications—from simple linear regression to multi-level/hierarchical models, time-series analysis, and regression—with multi-language interfaces for Python (CmdStanPy, legacy PyStan), R (CmdStanR, RStan), Julia (Stan.jl), MATLAB, Mathematica, and Unix shell, plus a browser-based Stan Playground hosted by the Flatiron Institute, high-level R packages (brms, rstanarm), and an extensive diagnostics/visualization ecosystem (bayesplot, ArviZ, loo, posterior, projpred, MCMC Monitor).

The project serves academic researchers, data scientists, statisticians, and students in industry and academia who need flexible, scalable Bayesian modeling and accurate uncertainty quantification. Stan's business model is non-commercial: the software is distributed free under the BSD 3-clause license (with documentation under CC-BY-ND 4.0), with revenue replaced by grants, donations, and corporate gifts channeled through NumFOCUS. Funders include the U.S. Department of Energy, U.S. National Science Foundation, U.S. National Institutes of Health, U.S. Office of Naval Research, U.S. Department of Education, Alfred P. Sloan Foundation, Chan Zuckerberg Foundation, Research Council of Finland, Simons Foundation (hosting/CI), Novartis, Google, and Meta. Governance rests with a three-person Stan Governing Body (Jesse Piburn, Teddy Groves, Aki Vehtari) representing Oak Ridge National Laboratory, Danish Technical University, and Aalto University / ELLIS Institute Finland respectively. Go-to-market is community-led: adoption is driven through the Stan Forums (Discourse), Slack, GitHub (stan-dev), documentation, tutorials, case studies, academic publications, and the annual StanCon conference series (StanCon 2026 scheduled for Uppsala, Sweden).

Name

Stan

Legal name

Stan

Website

https://mc-stan.org

Company type

Private

Founded year

2011

Operating status

Operating

Headcount range

11–50 employees

Short description

Stan is an open-source probabilistic programming language for Bayesian statistical inference, using Hamiltonian Monte Carlo and the No-U-Turn Sampler, distributed free to academic researchers, data scientists, statisticians, and students through language interfaces for Python, R, Julia, MATLAB, and Mathematica.

Ownership category

akta.pro rank

Product category

Probabilistic Programming / Bayesian Statistical Computing

NAICS

Software Publishers (5132), Scientific Research and Development Services (5417)

SIC

Services-Prepackaged Software (7372)

akta.pro primary industry

Financial Planning & Analysis (FP&A) & BI (HDAEAMAG)

akta.pro secondary industries

Computerized Adaptive Testing (CAT) & Psychometric Analytics Platforms (EDAFADAD), Semantic Layer, Metrics Store & Headless BI (HDAEAMAE), Bioinformatics & Multi-omics Analysis Software (HLAGAJAD)

Keywords

• Probabilistic programming
• Bayesian inference
• Statistical modeling
• Monte carlo sampling
• Stan language

Headquarters

HQ city

Trois-rivières

HQ country

Canada

HQ region

North America

Offices1 record

Markets served

GTM type

B2C

Offering type

Software

Cost components

Technology or R&D, Operations, Marketing or Sales, Infrastructure

Revenue model

1. Open Source Software (Free): Stan software is freely available under BSD 3-clause license. Development is funded through grants, corporate donations, and nonprofit support rather than software sales.

Pricing tiers

Go-to-market motion1 record

Distribution channels6 records

Marketing channels6 records

Core offering

Stan is an open-source probabilistic programming language and ecosystem for Bayesian statistical inference. The platform uses Hamiltonian Monte Carlo (HMC) and the No-U-Turn Sampler (NUTS) to fit statistical models specified by the user, supporting applications ranging from simple linear regression to complex multi-level hierarchical models and time-series analysis. It is distributed for free under the BSD 3-clause license and exposes interfaces for Python, R, Julia, MATLAB, Mathematica, and Unix shell, accompanied by a rich ecosystem of diagnostics and visualization packages.

Product overview

Stan is an open-source probabilistic programming language and ecosystem for Bayesian statistical inference. The core product consists of the Stan language and compiler (stanc3), the CmdStan command-line interface, and language-specific interfaces for Python (CmdStanPy), R (CmdStanR), Julia (Stan.jl), MATLAB, and Mathematica. The ecosystem includes high-level modeling packages like brms and rstanarm that provide formula-based interfaces without requiring users to write Stan code directly. Supporting tools include the Stan Math Library (C++ automatic differentiation backend), BridgeStan (multi-language bindings), visualization packages (bayesplot, ArviZ), and diagnostic tools (loo, posterior, projpred, MCMC Monitor). StanCon is the annual community conference. Stan uses Hamiltonian Monte Carlo (HMC) and No-U-Turn Sampler (NUTS) algorithms for posterior inference, supporting applications from simple linear regression to complex hierarchical and time-series models.

Differentiator

Problem solved

Functional benefit

Products and services

• Stan (Core Platform) Open-source probabilistic programming language for Bayesian statistical inference using Hamiltonian Monte Carlo (HMC) and No-U-Turn Sampler (NUTS), enabling sophisticated statistical modeling from simple regression to complex multi-level and time-series models for researchers and data scientists.
• CmdStan Command-line interface to Stan, usable from any shell environment, providing the core compilation and execution engine that language-specific interfaces wrap for developers and advanced users.
• CmdStanPy Recommended Python interface to Stan based on CmdStan, allowing Python users to write Stan models, compile them, and run inference directly from Python for data scientists and researchers.
• CmdStanR Recommended R interface to Stan based on CmdStan, providing seamless integration for R users to specify models and perform Bayesian inference within the R statistical computing environment.
• Stan.jl Interface to Stan for Julia users based on CmdStan, enabling Julia programmers to leverage Stan's inference engine within the Julia scientific computing ecosystem.
• Stan Playground Browser-based editor and runtime environment for Stan models, hosted by the Flatiron Institute, allowing new and existing users to try Stan without local installation.
• Stan Math Library C++ template library providing automatic differentiation and mathematical functions that serve as the computational backend for Stan's gradient-based inference algorithms.
• stanc3 Compiler Stan compiler written in OCaml that translates Stan programs (`.stan` files) into C++ code for subsequent compilation and execution by the Stan Math Library.
• BridgeStan Library providing bindings to a Stan model's log densities, gradients, and Hessians for C++, Python, Julia, R, and Rust, enabling developers to embed Stan models within other applications.

Segments3 records

Ideal customer profiles3 records

Technology focussed Yes

API detail

Has API

No

API docs

API detail

Core technology

AI maturity

App detail

Integration19 records

AI capability3 records

Feature8 records

Partnerships

One partnership is on record.

• Flatiron InstitutemajorStrategic or Co-development PartnerFlatiron Institute hosts Stan Playground, a browser-based editor and runtime environment for Stan models. This is a key web-based tool recommended for new users.

Scale indicators4 records

Recent moves5 records

Expansion highlights5 records

Others

• Statistical Rethinking (Richard McElreath): Educational ecosystem centered on Bayesian modeling using Stan via the 'rethinking' R package. Ecosystem/adjacent peer that channels large cohorts of instructors and students into Stan and amplifies its teaching footprint.

Direct peers

• Pyro: Uber-developed probabilistic programming language built on PyTorch, supporting Bayesian inference and deep generative models. Direct competitor that targets a more deep-learning-native workflow than Stan.
• Turing.jl: Julia-native probabilistic programming library for Bayesian inference, providing similar modeling abstractions to Stan but inside the Julia ecosystem. Directly comparable on algorithmic approach and target user persona.
• JAGS: Long-standing open-source Gibbs sampler for Bayesian hierarchical models, often cited alongside Stan as a primary alternative for MCMC-based Bayesian inference in academic settings.
• PyMC: Open-source probabilistic programming framework for Bayesian modeling in Python, using Hamiltonian Monte Carlo and NUTS. Direct competitor to Stan for Python-first statistical inference, with overlapping users in academia and industry.
• OpenBUGS: Open-source Bayesian inference software using BUGS-language model specification. Direct historical peer to Stan that has been progressively displaced in many advanced use cases by Stan's HMC/NUTS algorithms.
• NumPyro: JAX-based probabilistic programming library offering scalable Bayesian inference with HMC and NUTS. Direct competitor to Stan optimized for GPU and large-scale models, sharing the same underlying algorithmic lineage.

Broad incumbents

• TensorFlow Probability: Google-maintained library adding probabilistic modeling and Bayesian inference to the TensorFlow ecosystem. Broader incumbent that competes with Stan for industrial and deep-learning-anchored Bayesian use cases.
• IBM SPSS Statistics: Commercial statistical software package used by enterprises and academic researchers for advanced modeling. Adjacent competitor in the broader statistical modeling category where Stan operates.
• SAS/STAT: Commercial statistical analysis suite widely used in enterprise and government for regression, Bayesian, and mixed-model workloads. Overlaps with Stan on advanced statistical modeling, especially in regulated industries.

Market position

Strengths4 records

Weaknesses3 records

Competitive moat5 records

Key risks6 records

Key highlights6 records

Customer concentration

• Xx.com/mcmc_stan
• GitHubgithub.com/stan-dev
• LinkedInlinkedin.com/company/stancon-2020-oregon-state-university

Compliance4 records

Revenue estimate

Valuation estimate

Number of profiles

Profiles3 records

Funding overview

Funding rounds

Investors

Funding detail is available on the Subscription and Enterprise plan.Contact sales →

M&A

Investments

M&A and investment is available on the Subscription and Enterprise plan.Contact sales →

Frequently asked questions about Stan