Build, run, and share computational pipelines with drag-and-drop nodes. Molecular dynamics, structure prediction, docking: on your own machine, your cluster, or managed GPUs.
v0.17.3 · macOS (Apple Silicon & Intel) · Linux · Windows · Free
Select a waypoint to read its name
The route above draws a real template, PDBmdAuto, from the pdbmdauto package on Salpa Hub. Give it a PDB code and it rebuilds the residues the experiment did not resolve, sets protonation states, solvates and relaxes the system, then runs a short GROMACS simulation, keeping every intermediate file.
Drag real science tools onto a canvas and wire them into a pipeline.
On your laptop or your cluster. GPU models such as Boltz-2 run on Salpa Compute.
Export it as one JSON file with every node and setting. Keep it with your data, or send it to a colleague.
Open a real pipeline
A 15-node metallopeptide MD workflow — node library, canvas, and execution panels in one pass.
Salpa, Salpa Hub, Salpa Compute and Salpa CLI.
The desktop app
Open workflow packages
hello-world-pipeline, the quickest way to see how nodes wire togetherAI models as cloud nodes
Boltz-2 structure predictionChai-1 structure prediction of complexesDiffDock blind dockingEvo2 genomicsProteinMPNN protein designMore capacity by arrangement; nothing is sold self-service
Node packages from the terminal
pip install salpa-clinew scaffold a packagevalidate check it installs and registerssmoke run a node on its sample datapush put it into the app on your machinemetalparm-vwfTake a metal complex to a QM-derived AMBER force field and an MD-ready topology. Parameterise a cofactor once, reuse it across peptide sequences.
20 nodes · LGPL-2.1metaldock-vwfDock metal complexes: receptor and ligand preparation, metal-aware QM charges, then docking.
6 nodes · Non-commercialpdbmdautoA PDB code in, an MD-ready system out. The route at the top of this page is its template.
13 nodes · MIThello-world-pipelineThe quickest way to see how nodes wire together: encrypt a message, decrypt it, show both.
3 nodes · MITsmiles-to-3d3D conformers from a SMILES string, written as SDF, ready for docking or MD.
1 node · BSD-3-Clauseligand-pdbqt-meekoDocking-ready PDBQT ligands from an SDF, prepared with Meeko.
1 node · LGPL-2.1Install them inside the app: Marketplace, where Salpa Hub is the built-in source.github.com/BoundaryComputing/salpa-hub
Salpa CLI
Scaffold a package, check it, run a node on its sample data, then put it into the app on your machine. Nothing is uploaded.
Publishing to Salpa Hub is done by us, after review.
The Salpa CLI pageaddenvplatformsunpushdocsFree for macOS (Apple Silicon and Intel), Linux and Windows. Current version 0.17.3.
Get started guideWant a hand with your own work? Salpa Foundry gives free, hands-on help.