Visual workflows
for scientific computing
Build, run, and share computational pipelines with drag-and-drop nodes. Molecular dynamics, structure prediction, docking — on your machine, with cloud GPUs when you need them.
v0.3.1 · macOS (Apple Silicon & Intel) · Linux · Windows · Free
PDBmdAuto — an 11-node pipeline from PDB structure to GROMACS MD, on the Salpa canvas
Built-in & cloud nodes for
Empowering workflows,
not technical burdens
Computational science has powerful tools — but wiring them together remains a technical challenge. Here is how the current landscape looks.
Web Platforms
e.g. Galaxy
- User-friendly interface
- Managed infrastructure
- Data must leave your machine
- Limited to pre-installed tools
Script Pipelines
e.g. Nextflow, Snakemake
- Powerful and flexible
- Reproducible environments
- DSL required — steep learning curve
- Limited adoption by non-programmers
Wrapped Packages
e.g. Python packages with config files
- Familiar Python ecosystem
- Convenient for single tools
- Workflow features are an afterthought
- Extending means reverse-engineering
Open science solved access, data, source, and protocols. Salpa targets the remaining gap — making workflows approachable for both developers and domain scientists.
Everything a pipeline needs, in the box
Visual Workflow Canvas
Drag nodes onto a canvas and wire them together. What Nextflow and Snakemake ask for in DSL code, Salpa does with drag and drop — accessible to every scientist, not just programmers.
Real capture — the PDBmdAuto template on the canvas
Pure-Python Nodes
Your script becomes a visual node with an auto-generated panel. No framework, no server, no XML.
Zero-Config Environments
Each node resolves its own isolated environment. No pip conflicts, no Docker.
Local · Cloud · HPC
One interface for every execution target, with live progress.
Your Workflow Is a File
Every pipeline exports as a plain, portable file — share it, rerun it anywhere, attach it to a paper. Reproducibility is the default, not an afterthought.
Marketplace & Hub
Install community nodes from the open Salpa Hub, GitHub, or local folders. No lock-in.
Real capture — the library after a Hub install
Data Sovereignty
Your data stays on your machine — always. Cloud runs are opt-in, ephemeral, no retention.
Five principles, chained like a pipeline
Simple
Complexity absorbed, simplicity delivered.
Workflows by drag and drop. Nodes in pure Python.
Open
See, change, and share how it works.
Open-source nodes and workflows, community-driven.
Local
Privacy by default, not by request.
Your machine is the hub; results flow back to you.
Visual
See your pipeline, not just read it.
Every connection, parameter, and result on the canvas.
Extensible
Integration over reinvention.
Wrap any tool or service into a shareable node.
See Salpa in action
Drag nodes onto the canvas, connect them, configure each step in a panel, and execute with one click — live logs included.
Open a real pipeline
A 15-node metallopeptide MD workflow — node library, canvas, and execution panels in one pass.
Free GPU computing
with a free account
Run deep-learning models that need heavy GPUs and complex environments — directly from Salpa. Register with your email or Google account and start computing.
More cloud nodes coming. Need a specific model? Suggest a tool — we deploy fast.
Ready to run your first workflow?
Free on macOS, Linux, and Windows — your first pipeline runs in minutes.
Salpa is in active development with early users — your feedback shapes what we build next. Early adopters get:
Fast Bug Fixes
Report issues and get fixes within days, not months.
Node Dev Help
We help you build custom computational nodes for your research.
HPC Integration
Connect Salpa to your institution's SLURM clusters.
Feature Requests
Suggest workflows, cloud nodes, and tools. Your ideas shape the roadmap.
Reach out anytime
hello@salpa.app