# Mamba — Fast Cross-Platform Package Manager

> A drop-in conda replacement written in C++ that resolves environments in seconds instead of minutes.

## Install

Save as a script file and run:

# Mamba — Fast Cross-Platform Package Manager

## Quick Use
```bash
# install micromamba (standalone, no base env needed):
curl -Ls https://micro.mamba.pm/install.sh | bash
# create an environment:
micromamba create -n myenv python=3.12 numpy pandas -c conda-forge
micromamba activate myenv
```

## Introduction
Mamba is a reimplementation of the conda package manager in C++ using the libsolv dependency solver. It provides the same CLI interface and channel compatibility as conda while delivering 10-50x faster environment resolution and parallel package downloads.

## What Mamba Does
- Resolves package dependencies using the SAT-based libsolv library for near-instant solves
- Downloads packages in parallel with multi-threaded HTTP fetches
- Creates, updates, and removes conda-compatible environments
- Reads environment.yml and conda-lock files for reproducible setups
- Ships as micromamba — a statically linked single binary with no Python dependency

## Architecture Overview
Mamba replaces conda's Python solver with libsolv, a C library originally built for RPM/APT that uses Boolean satisfiability to resolve dependencies. Package metadata is fetched and cached in a binary repodata format for fast reloads. Downloads use libcurl with connection pooling and parallel transfers. micromamba embeds the entire stack in one static binary, eliminating the need for a base conda installation.

## Self-Hosting & Configuration
- Install micromamba via the official shell script or Homebrew
- Use `micromamba create -n env -c conda-forge` to create environments
- Configure default channels and settings in `~/.mambarc` or `~/.condarc`
- Export environments with `micromamba env export` for reproducibility
- Integrates with conda-lock for deterministic cross-platform lockfiles

## Key Features
- 10-50x faster dependency resolution than conda
- Parallel multi-threaded package downloads
- Full conda compatibility — same channels, same package format
- micromamba: single static binary, no Python runtime needed
- Supports environment.yml, conda-lock, and explicit spec files

## Comparison with Similar Tools
- **conda** — the original; same ecosystem but significantly slower solver and serial downloads
- **uv** — ultra-fast pip replacement for PyPI packages; does not support conda channels
- **pixi** — Rust-based conda-compatible manager with project-level lockfiles and task runner
- **poetry** — Python dependency manager for PyPI; no support for conda-forge or non-Python packages
- **Spack** — HPC-focused package manager that builds from source; different target audience

## FAQ
**Q: Is mamba a full replacement for conda?**
A: Yes. Mamba and micromamba implement the same CLI commands and read the same configuration files. You can switch without changing workflows.

**Q: What is micromamba?**
A: A standalone statically linked binary that provides the mamba CLI without requiring a base conda installation or Python interpreter.

**Q: Does mamba work with conda-forge?**
A: Yes. It uses the same channel infrastructure and repodata format as conda.

**Q: Can I use mamba in CI/CD pipelines?**
A: Yes. micromamba is ideal for CI because it installs in seconds and has no bootstrapping dependency.

## Sources
- https://github.com/mamba-org/mamba
- https://mamba.readthedocs.io

---
Source: https://tokrepo.com/en/workflows/aac1a89a-4211-11f1-9bc6-00163e2b0d79
Author: Script Depot