Why Operon?
Operon is purpose-built for bioinformaticians who want a modern, AI-augmented development environment without leaving their domain tools behind.
Built for Biology
Understands bioinformatics file formats (FASTA, FASTQ, VCF, BAM, GFF), common pipelines, and domain-specific best practices out of the box.
Three AI Modes
Agent mode executes multi-step tasks. Plan mode architects solutions. Ask mode answers questions — with optional PubMed integration.
Remote HPC
Connect to university clusters and remote servers via SSH. Edit, run, and debug code on remote machines as if they were local.
Analysis Protocols
Curate and share reusable analysis protocols. Generate new ones with AI or write them manually in Markdown.
Git Integration
Full Git and GitHub workflow built into the sidebar — stage, commit, push, and publish repositories without leaving the app.
Native Performance
Built with Tauri (Rust) and React. Lightweight, fast, and uses your system's native webview — not Electron.
Installation
Operon is distributed as a native macOS application. Download the .dmg file and drag it into your Applications folder.
Download & Open the DMG
Double-click the downloaded .dmg file. You'll see the Operon icon ready to drag into Applications.
Setup Wizard
When you first launch Operon, a guided setup wizard walks you through installing all required dependencies. The process has three phases.
Phase 1: Xcode Command Line Tools
Xcode CLT provides essential build tools (git, make, clang) needed by almost everything else. Click Install Xcode CLI to begin.
A system dialog will appear asking you to confirm the installation. Accept the license agreement and wait for the download to complete.
xcode-select --install
Phase 2: Developer Tools
This phase installs Homebrew (the macOS package manager), Node.js, and GitHub CLI. Click Install Developer Tools and enter your password when prompted.
# Homebrew
/bin/bash -c "$(curl -fsSL https://raw.githubusercontent.com/Homebrew/install/HEAD/install.sh)"
# Node.js
brew install node
# GitHub CLI
brew install ghPhase 3: Claude Code
Operon uses Claude Code (Anthropic's CLI agent) as its AI backend. This step installs it globally via npm.
npm install -g @anthropic-ai/claude-code
Authentication
Operon needs access to Claude AI. You can authenticate via OAuth (recommended — uses your Anthropic account) or by providing an API key.
OAuth Login (Recommended)
Click Open Login Page. Operon will launch Claude's authorization page in your browser. Sign in, authorize the app, and paste the resulting code back into Operon.
API Key
Alternatively, paste an Anthropic API key directly. This is stored securely in the system keychain. Useful for institutional or shared setups.
App Tour
After authentication, Operon walks you through a quick visual tour highlighting the key areas of the interface.
Workspace Overview
The Operon workspace is divided into four main areas: the activity bar on the left, the sidebar, the editor / terminal in the center, and the AI chat panel on the right.
The activity bar (thin left column) lets you switch between views: File Explorer, Git, SSH, Protocols, Settings, and Help. The main area adapts based on which view is selected.
File Explorer
Browse, create, rename, and delete files and folders directly within Operon. The explorer supports project indexing for quick navigation.
Right-click any file or folder for context actions. The file explorer also serves as the entry point for opening files in the editor. Supported file types get syntax highlighting automatically.
Code Editor
Operon includes a built-in code editor with syntax highlighting for all major bioinformatics languages: Python, R, Bash, Nextflow, Snakemake, and more.
Files open in tabs in the main panel. The editor supports standard features like find-and-replace, line numbers, word wrap, and configurable font size. You can adjust all of these in Settings.
.py, .R, .sh, .nf (Nextflow), .smk (Snakemake), .wdl (WDL), .json, .yaml, .toml, .md, .csv, .tsv, and many more.
Integrated Terminal
A fully functional terminal is embedded at the bottom of the workspace. It uses your system shell (zsh/bash) and inherits your existing environment, aliases, and conda/mamba setups.
You can resize the terminal by dragging its top edge, or toggle it with ⌘ + `. Multiple terminal sessions are supported. The terminal is also used by Claude in Agent mode to execute commands on your behalf.
~/.zshrc and ~/.bashrc — conda environments, custom PATHs, and module systems work seamlessly.
AI Chat Panel
The right-hand panel is your direct line to Claude. Type natural-language requests and Claude will respond with explanations, code, or take actions in your workspace.
The chat panel supports full Markdown rendering, code blocks with syntax highlighting, and inline action buttons. You can start new sessions, rename them, or revisit previous conversations from the session list.
Three AI Modes
Switch between modes using the dropdown at the top of the chat panel. Each mode changes how Claude approaches your request.
Agent Mode
Agent is the default and most powerful mode. Claude can read and write files, run terminal commands, install packages, execute scripts, and iteratively debug — all autonomously. Use this for tasks like "Run a DESeq2 analysis on my count matrix" or "Set up a Nextflow pipeline for variant calling."
Plan Mode
Plan mode asks Claude to think before it acts. Instead of executing immediately, Claude produces a detailed step-by-step plan that you can review and approve before execution. Ideal for complex multi-stage workflows where you want oversight.
Ask Mode
Ask mode is pure Q&A. Claude answers questions, explains code, and provides guidance — without making any changes to your files or running any commands. Perfect for learning, troubleshooting, or literature review.
Analysis Protocols
Protocols are reusable, domain-specific instructions that guide Claude's behavior for particular types of analyses. Think of them as expert templates — they encode best practices for tools, data formats, quality control, and reproducibility.
Operon ships with a comprehensive library of over 180+ built-in protocols covering RNA-seq, ChIP-seq, ATAC-seq, WGS, scRNA-seq, metagenomics, GWAS, proteomics, molecular dynamics, and more. When you select a protocol and start a session, its instructions are automatically injected into Claude's context.
Creating Your Own Protocols
You're not limited to the built-in set. Click the + button in the protocols sidebar to create your own. There are two ways:
AI-Generated (Recommended)
Describe what you need in plain English — for example, "A protocol for 16S rRNA amplicon analysis using QIIME2 with DADA2 denoising" — and Claude will generate a comprehensive protocol with sections for environment setup, SLURM configuration, tool parameters, QC checkpoints, and reproducibility guidelines.
Manual Editor
Prefer to write your own? Switch to the "Write Manually" tab and compose your protocol in Markdown. You have full control over the structure and content.
~/.operon/protocols/. You can edit them directly with any text editor, share them with colleagues, or version-control them with Git.
PubMed Search
Operon integrates with NCBI's PubMed database so Claude can search and cite biomedical literature directly in your conversations.
Enable the PubMed toggle (available in Ask mode) and ask questions like "What are the latest methods for single-cell ATAC-seq peak calling?" Claude will search PubMed, retrieve relevant papers, and incorporate their findings into its answer with proper citations.
This is powered by the NCBI E-utilities API and requires no additional setup or API keys.
Tutorial: Running a Local Analysis
A practical, step-by-step walkthrough for first-time users. By the end of this tutorial you will have run a complete analysis on your own machine using Operon.
Install Operon
Download the correct installer for your Mac from the buttons at the top of this page (Apple Silicon or Intel). Open the .dmg and drag Operon into your Applications folder. On first launch the Setup Wizard will install all required dependencies — just follow the prompts.
Authenticate with Claude
When prompted, log in with your Anthropic account or paste your API key. Operon needs this to communicate with Claude. You can change your key later in Settings.
Open Your Data Folder
Click File → Open Folder (or use ⌘ + O) and navigate to the folder that contains your data — for example, a directory with FASTQ files, a count matrix, or VCF files. This sets the working directory so Claude can see and access your files.
Select a Protocol
Click the Protocols icon in the activity bar and browse the library of 180+ built-in protocols. Pick the one that matches your analysis — for example, DESeq2 Differential Expression or scRNA-seq (Seurat). Selecting a protocol gives Claude expert-level context for that specific workflow.
Start in Plan Mode — Iterate Before Executing
Switch the mode dropdown to Plan and describe what you want to do — for example, "I have paired-end RNA-seq FASTQs in this folder. Run QC, align to GRCh38, and perform differential expression with DESeq2."
Claude will produce a detailed, step-by-step plan — the tools it will use, the order of operations, expected outputs, and any assumptions. Read the plan carefully. Ask follow-up questions, request changes, or point out things you want done differently. Keep iterating in Plan mode until you are fully confident the plan is correct.
Execute with Agent Mode
Once you are happy with the plan, switch to Agent mode. You can paste or reference the finalized plan and tell Claude to proceed. Agent mode will execute each step autonomously — installing packages, writing scripts, running commands, and handling errors — while you watch the progress in the chat and terminal panels.
Use Ask Mode for Quick Questions
At any point during your analysis, switch to Ask mode to get explanations without Claude making any changes. This is useful for understanding outputs ("What does this volcano plot tell me?"), clarifying statistics ("Why did DESeq2 shrink these log-fold changes?"), or searching the literature with PubMed toggled on.
Review Results and Iterate
Check the output files in the File Explorer. If you need to adjust parameters, add a visualization, or run a follow-up analysis, just tell Claude. You can go back and forth between Plan and Agent modes as many times as you need. Each session preserves the full conversation history so Claude remembers what was done previously.
Tutorial: Running on a Remote Server / HPC
Most bioinformatics workloads run on institutional HPC clusters or cloud servers. This tutorial walks you through using Operon to work on a remote machine, with practical tips for long-running jobs.
Install Operon Locally
Install Operon on your Mac using the same steps from the local tutorial above. Operon runs locally on your laptop and connects to the remote server over SSH — you do not need to install Operon on the server itself.
Connect via SSH
Click the SSH icon in the activity bar and set up a new connection with your server's hostname, port, and credentials. Operon supports password, SSH key, and Duo/MFA authentication. Once connected, the file explorer and terminal switch to the remote filesystem.
Start a tmux or screen Session
Before doing anything else, open the terminal panel and start a persistent terminal multiplexer session:
# Using tmux (recommended)
tmux new -s operon
# Or using screen
screen -S operonThis is critical for remote work. If your SSH connection drops — laptop goes to sleep, Wi-Fi hiccup, VPN reconnect — a bare SSH session would kill any running processes. With tmux or screen, your session survives disconnections and you can reattach later:
# Reattach to tmux
tmux attach -t operon
# Reattach to screen
screen -r operonRequest an Interactive Compute Node
On most HPC clusters the login node is shared and has limited resources. Do not run analyses on the login node. Instead, request an interactive session on a compute node:
# SLURM (most common)
srun --pty --mem=16G --cpus-per-task=4 --time=4:00:00 bash
# PBS/Torque
qsub -I -l nodes=1:ppn=4,mem=16gb,walltime=04:00:00
# SGE
qlogin -l h_vmem=16G,h_rt=04:00:00Once you land on a compute node, Claude can use it to run tools, process data, and execute scripts with the allocated resources. This is where your actual analysis will happen.
Navigate to Your Data
Use the file explorer or terminal to cd into the directory where your data lives — for example, /scratch/username/rnaseq_project/. Open this folder in Operon so Claude can see the files.
Select a Protocol and Plan First
Just like the local workflow: pick a protocol, switch to Plan mode, and describe your analysis. Iterate on the plan until you are confident. This is even more important on a remote server where resources are limited and mistakes can waste queued compute time.
When discussing the plan, tell Claude about your cluster environment — available modules (module avail), conda environments, scratch vs. home directories, and any job submission conventions your lab follows.
Execute in Agent Mode
Switch to Agent mode and let Claude run the analysis on the compute node. Claude will write scripts, execute commands, and adapt based on output — all within your tmux/screen session. For very long-running steps (e.g., genome alignment), Claude can write and submit a batch job script instead of running interactively:
# Claude can generate and submit SLURM scripts like:
sbatch alignment_job.shYou can then monitor the job with squeue -u $USER or ask Claude to check on it.
Reconnect If Disconnected
If your connection drops, simply reconnect via SSH in Operon and reattach your tmux/screen session. Your analysis will still be running (or will have completed) and all the output will be intact. Start a new Operon chat session and point Claude to the working directory — it can pick up where things left off by examining the files and logs.
SSH Connections
Connect to remote servers and HPC clusters directly from Operon. Browse files, edit code, and run analyses on remote machines without leaving the app.
Setting Up a Connection
Click the SSH icon in the activity bar, then New Connection. Fill in your server details: hostname, port (defaults to 22), username. Operon supports both password and SSH key authentication, as well as Duo/MFA for university clusters that require two-factor auth.
Working Remotely
Once connected, the file explorer switches to show the remote filesystem. You can browse directories, open files in the editor, and run commands in a remote terminal session — all through the same interface you use locally.
ssh-keygen and ssh-copy-id.
Remote Workflows
Operon can install Claude Code on your remote server and run AI-assisted sessions directly on HPC infrastructure.
This means you can use Agent mode on a remote cluster: Claude will execute commands via SSH, write scripts to the remote filesystem, submit SLURM/PBS jobs, and monitor their progress. Your data never leaves the server — only the terminal I/O travels over the SSH connection.
The setup wizard on the remote tab checks for Node.js and Claude Code on the remote machine and offers to install them if missing.
Git & GitHub
Operon includes a full Git integration in the sidebar. Track changes, commit, push, and publish repositories — all without leaving the app.
Key Git Features
Initialize a Repository
If your project doesn't have a .git folder, Operon shows an Initialize Repository button. Click it to run git init and start tracking your files.
Stage, Commit & Push
The Git panel lists all modified, added, and deleted files. Write a commit message and click Commit All to stage and commit everything at once. Then click Push to send your changes to the remote.
Publish to GitHub
Authenticate with GitHub via gh auth login (Operon walks you through this). Then use the Publish button to create a new GitHub repository and push your code — public or private, your choice.
Version Tagging
Operon supports semantic version tagging. Tag releases directly from the Git panel to mark milestones in your analysis pipelines.
Settings
Customize Operon to match your preferences. Access settings from the gear icon in the activity bar.
Available Settings
| Setting | Description |
|---|---|
| Font Size | Adjust the editor and terminal font size. |
| Font Family | Choose your preferred monospace font. |
| Tab Size | Set tab width (2 or 4 spaces). |
| Word Wrap | Toggle soft line wrapping in the editor. |
| Line Numbers | Show or hide line numbers. |
| Minimap | Toggle the code minimap on the right edge. |
| Auto-save | Automatically save files after changes. |
| Theme | Switch between dark themes. |
| Voice Dictation | Enable voice input for the AI chat panel. |
Keyboard Shortcuts
Speed up your workflow with these keyboard shortcuts.
| Shortcut | Action |
|---|---|
| ⌘ + S | Save current file |
| ⌘ + P | Quick file search |
| ⌘ + ` | Toggle terminal |
| ⌘ + B | Toggle sidebar |
| ⌘ + J | Toggle bottom panel |
| ⌘ + N | New file |
| ⌘ + ⇧ + N | New folder |
| ⌘ + W | Close current tab |
| ⌘ + Enter | Send message in AI chat |
| ⌘ + / | Toggle comment in editor |
| ⌘ + F | Find in file |
| ⌘ + ⇧ + F | Find in project |
| Esc | Stop AI session / Cancel |
Help Panel
The built-in help panel (click the ? icon in the activity bar, or use the Help menu) provides quick-access documentation organized into sections.
The help panel covers:
Getting Started
Overview of the workspace layout, how to open projects, and first steps with the AI assistant.
Working with Files
How to create, rename, delete, and organize files. Project indexing for large codebases.
Using the Terminal
Terminal basics, resizing, multiple sessions, and environment configuration.
AI Assistant
Deep dive into Agent, Plan, and Ask modes. How to write effective prompts. Session management.
Analysis Protocols
How protocols work, using built-in protocols, and creating your own (AI-generated or manual).
Remote Servers (SSH)
Setting up connections, SSH key management, Duo/MFA, and running remote sessions.
Git & GitHub
Repository management, committing, pushing, and publishing to GitHub.
PubMed Integration
Searching biomedical literature within the AI chat interface.
Voice Dictation
Enable and use voice input for hands-free interaction with the AI assistant. Powered by macOS dictation.
Customization
Theme settings, font configuration, editor preferences, and workspace customization.
Keyboard Shortcuts
Complete list of keyboard shortcuts for navigation, editing, and AI interaction.
Troubleshooting
Common issues with installation, authentication, SSH connections, and Claude Code — plus their fixes.
Frequently Asked Questions
What is Claude Code and why does Operon need it?
Claude Code is Anthropic's command-line AI agent. Operon uses it as the backend for all AI features — it's what allows Claude to read your files, run commands, and execute multi-step analyses. Operon provides the visual interface; Claude Code provides the AI engine. For more details, see the Claude Code documentation.
Does my data leave my computer?
When using Claude, your prompts and relevant file contents are sent to Anthropic's API for processing. For remote/HPC workflows, data stays on the remote server — only terminal I/O is transmitted over SSH. Operon itself doesn't collect or transmit any telemetry.
Can I use Operon offline?
The editor, file explorer, terminal, and Git features all work offline. AI features (Agent, Plan, Ask, PubMed) require an internet connection to reach the Claude API.
What HPC schedulers are supported?
Operon's protocols include templates for SLURM, PBS/Torque, and SGE. Claude can generate and submit job scripts for any of these schedulers when working in Agent mode on a remote cluster.
Can I use my own conda environments?
Yes. The integrated terminal inherits your shell configuration, including conda/mamba initialization. Your environments are available in both the terminal and when Claude executes commands in Agent mode.
How do I update Operon?
Download the latest .dmg from the releases page and replace the existing application in your Applications folder. Your settings, protocols, and sessions are stored in ~/.operon/ and won't be affected.
Is Operon open source?
Operon is built with Tauri (Rust) and React. Check the project repository for licensing details and contribution guidelines.
System Requirements
| Requirement | Details |
|---|---|
| Operating System | macOS 12 (Monterey) or later |
| Architecture | Apple Silicon (M1/M2/M3/M4) or Intel — separate downloads available for each |
| Disk Space | ~500 MB (including dependencies) |
| RAM | 4 GB minimum, 8 GB recommended |
| Internet | Required for AI features and initial setup |
| Dependencies | Installed automatically: Xcode CLT, Homebrew, Node.js, GitHub CLI, Claude Code |