Free Molecular Docking Software in 2026: The Complete List

AutoDock Vina is the default choice for academic molecular docking, and for good reason. It is fast (10–1000× faster than its predecessor AutoDock 4), free, open source, and well-documented. The overwhelming majority of academic docking papers published in the last decade used Vina or a Vina-derived tool. If you’re learning docking for the first time, this is where you start.

Vina uses an iterated local search global optimizer to explore conformational space and a hybrid scoring function combining empirical and knowledge-based terms. The 2021 release (v1.2) added GPU acceleration via AutoDock-GPU and support for additional scoring functions including Vinardo and the original AutoDock4 force field — making it more flexible and more competitive with newer tools.

How to install

The easiest installation on Linux and macOS is via conda:

conda install -c conda-forge autodock-vina

Or download the pre-compiled binary directly from the GitHub releases page and add it to your PATH. On Windows, download the .exe from the same releases page. Verify the install with:

vina --version

GNINA (pronounced “ninja”) is a Vina fork that replaces the traditional scoring function with a convolutional neural network trained on tens of millions of protein-ligand complexes from the PDB. That single change — same search algorithm, smarter scorer — consistently produces better pose prediction than standard Vina across virtually every benchmark it’s been tested on.

The practical implication: if you’re already comfortable running Vina, switching to GNINA is a near-zero-cost upgrade. The command-line interface is nearly identical, the input file formats are the same, and the output is the same PDBQT format you already know how to visualize. The only meaningful difference is that GNINA runs best on a GPU, and the binary is larger.

For projects where pose accuracy matters — hit validation, understanding binding mechanisms, structure-activity relationship analysis — GNINA is the best free option available. The accuracy improvement over Vina is real, reproducible, and well-documented in the published benchmarks.

How to install

The simplest route on Linux is to download the pre-compiled binary from the GitHub releases page — no compilation required:

wget https://github.com/gnina/gnina/releases/latest/download/gnina
chmod +x gnina
./gnina --version

On a system without a CUDA GPU, GNINA falls back to CPU mode automatically, though it will be slower than Vina. For Windows users, install via WSL2 (Windows Subsystem for Linux) and follow the Linux instructions.

SwissDock is a web server — you upload your protein and ligand files through a browser, configure a few options, submit the job, and receive results by email. No installation, no command line, no environment setup. For researchers who need a quick docking result without committing to a full software setup, it’s genuinely useful.

Under the hood, SwissDock uses EADock DSS, a docking algorithm developed at the University of Geneva. It performs a two-stage search: a fast initial clustering step to identify promising regions of the binding site, followed by a more thorough local optimization. Results are returned as a ranked list of clusters, viewable through an integrated 3D viewer.

The limitations are real. SwissDock runs on a shared server, so jobs queue and can take hours during busy periods. You’re limited in the size and number of jobs you can submit. And because you don’t control the software version or parameters, reproducibility can be tricky for publication. For exploratory work, it’s excellent. For virtual screening or anything requiring automation, it’s the wrong tool.

How to use

No installation needed. Go to swissdock.ch, upload your protein in PDB format and your ligand in MOL2 or SDF format, define the target region (or run a blind dock), and submit. Results arrive by email, typically within 1–6 hours depending on server load.

rDock was originally developed at GlaxoSmithKline for high-throughput virtual screening and later released as open source. It’s fast, highly configurable, and designed from the ground up for screening large libraries — making it a genuine alternative to Vina for virtual screening campaigns, particularly when you need fine-grained control over the docking protocol.

RxDock is the community-maintained fork that has picked up active development where rDock left off. It offers better compatibility with modern Linux systems, improved documentation, and continued bug fixes. If you’re choosing between the two today, use RxDock.

rDock/RxDock is particularly notable for its strong performance on nucleic acid targets — docking small molecules to RNA or DNA — where most other programs struggle. If your target is an RNA structure or a DNA G-quadruplex, rDock is one of the few free tools with established protocols for this use case.

How to install

RxDock is available via conda for the simplest install:

conda install -c conda-forge rxdock

After installation, set the required environment variable and verify:

export RBT_ROOT=$(conda info --base)/envs/your-env
rbdock --help

Note that rDock/RxDock requires a cavity definition file (.prm) and a pre-computed grid, which adds setup steps not present in Vina. The documentation at rxdock.gitlab.io walks through this in detail.

PLANTS uses ant colony optimization (ACO) — a nature-inspired algorithm modeled on how ants find shortest paths — to search conformational space. This gives it a fundamentally different search strategy from both Vina’s gradient-based approach and AutoDock’s genetic algorithm, and it performs particularly well in benchmarks for fragment-based docking and for cases where the binding site is shallow or open.

PLANTS outputs results with the CHEMPLP scoring function, which includes a hydrogen bond term, steric clash penalty, and metal coordination terms — making it a reasonable choice for metalloprotein targets that don’t involve true covalent coordination. It’s less widely used than Vina, which means fewer tutorials and a smaller community, but the software itself is mature and reliable.

How to install

PLANTS is distributed as a pre-compiled binary for Linux and macOS after free academic registration. Download the appropriate binary, make it executable, and add it to your PATH:

chmod +x PLANTS1.2_64bit
export PATH=$PATH:/path/to/plants/
PLANTS1.2_64bit --help

PLANTS uses a plain-text configuration file (plantsconfig) specifying the binding site center, radius, and input files — conceptually similar to Vina’s config file but with different syntax.