Install Autopipe App

Set up Autopipe in a few minutes and start building bioinformatics pipelines with AI.

1

Install Autopipe Desktop

Download the desktop app for your platform. It provides a GUI for configuration and runs as an MCP server for Claude Desktop.

macOS

Download the .dmg file from the latest GitHub release:

Download for macOS

Windows

Download the .msi installer from the latest GitHub release:

Download for Windows

Linux

Download the .deb or .AppImage from the latest GitHub release:

Download for Linux
2

Connect to Claude Desktop

Autopipe works as an MCP (Model Context Protocol) server inside Claude Desktop. Register it with one command:

autopipe --register

This automatically adds Autopipe to Claude Desktop's MCP configuration. Restart Claude Desktop after registering.

3

Configure Your Server

Open the Autopipe desktop app and set up your remote SSH server. This is where pipelines will be executed.

SSH Host Your compute server address
SSH User Username for SSH connection
Pipelines Directory Where pipeline files are stored on the server
Output Directory Where analysis results are saved

Make sure Docker is installed on your remote server. Pipelines run inside Docker containers for reproducibility.

4

Connect GitHub

Link your GitHub account in the desktop app's GitHub tab. This allows you to upload pipelines and publish them to AutoPipeHub.

A GitHub Personal Access Token with repo scope is required.

5

Create Your First Pipeline

Open Claude Desktop and describe what you want to analyze. For example:

Create a variant calling pipeline for paired-end WGS data using BWA-MEM2 and GATK HaplotypeCaller

Claude will use Autopipe to:

  1. Generate a Snakefile with the analysis workflow
  2. Create a Dockerfile with all required tools
  3. Write a config.yaml for your parameters
  4. Produce ro-crate-metadata.json for discoverability
  5. Add a README.md with usage instructions
6

Run & View Results

Once your pipeline is generated, ask Claude to execute it:

Build the Docker image and run a dry-run first, then execute on my samples in /data/wgs_samples

After execution, view results directly in the browser viewer or download them locally.

7

Share on AutoPipeHub

Publish your pipeline to make it available for others:

Upload this pipeline to GitHub and publish it to AutoPipeHub

Your pipeline will be searchable on AutoPipeHub and downloadable by anyone.

What's Next?

AutoPipeHub Browse and download community pipelines Plugins Extend the result viewer with custom plugins Plugin Guide Learn how to create and publish your own plugins