Mamba is a package manager that reimplements conda in C++ for faster performance across platforms.
Mamba addresses the speed limitations of conda by replacing key components with more efficient implementations. It uses parallel downloading with multi-threading to fetch repository data and packages concurrently, and replaces conda's dependency solver with libsolv, a state-of-the-art library used in Red Hat, Fedora, and OpenSUSE package managers. The core is written in C++ for maximum efficiency while maintaining compatibility by reusing conda's command-line parser, installation routines, and transaction verification code.
Mamba suits projects where installation speed matters, particularly in CI/CD pipelines and containerized environments. The project offers two variants: mamba itself, which integrates with existing conda environments and libraries, and micromamba, a statically linked standalone executable requiring no dependencies. Micromamba is preferred when a single self-contained binary is essential or runtime footprint must be minimal, while mamba is better when other software in the environment uses libmambapy or libmamba, or when regular library updates are needed. The tool adds a repoquery command for efficient repository and dependency queries, and micromamba can install lock files generated by conda-lock without requiring conda-lock itself.
Development activity shows consistent engagement with both maintenance and feature work. The project maintains active responsiveness to issues and pull requests across its codebase. Work spans the core C++ implementation, Python bindings, documentation updates, and integration tooling like the setup-micromamba GitHub Action. The team regularly addresses compatibility concerns and extends functionality while preserving the conda ecosystem integration that makes adoption straightforward for existing users.
