Skip to main content
Create reusable workflows in OpenBio to automate repetitive research tasks. Save sequences of steps as templates, then execute them with different inputs—perfect for standardizing analyses and saving time on routine tasks.
What is a Workflow? A workflow is a saved sequence of research steps that you can execute repeatedly. Think of it as a recipe for common analyses like “Protein Structure Prediction Pipeline” or “Literature Review Protocol.”It is a prompt that describes a sequence of steps to be executed. The more robust the workflow, the better the results will be.

​
Creating Workflows

​
From an Existing Chat

The easiest way to create a workflow is from a successful conversation:
1

Complete Your Analysis

Have a chat conversation where you perform the steps you want to save as a workflow
2

Ask OpenBio to save as workflow

Ask OpenBio to save as workflow. Again, it is very important to be specific and describe the workflow in detail.
You can also ask OpenBio to ask questions to you to help you describe the workflow in detail.
3

Review and Save

Review the workflow, ask OpenBio to make any changes, and save it will be saved to your workflow library automatically.
Your workflow is now available in the Workflows panel and can be referenced in any chat using /workflow-name

​
Creating from the Workflows Page

Build a new workflow without starting from a chat:
1

Go to Workflows Page

Click “Workflows” in the settings button in the top navigation bar
2

Create New Workflow

Click the ”+ Add Workflow” button in the top-right corner of the Workflows page
3

Design Your Workflow

  • Name: Give your workflow a descriptive title.
  • Description: Explain what it does and when to use it. Here’s what makes a good description:
    • Steps: Add steps manually with clear instructions.
    • Parameters: Define input parameters that will change each time.
    • Outputs: Set up expected outputs and results.
    • Helpful Descriptions: Add helpful descriptions for each step.
4

Test and Refine

Test your workflow with sample inputs to ensure it works correctly

​
Using Workflows

Reference workflows directly in your messages: 
Run /protein-analysis on this sequence: MKTAYIAKQRQ...
You can also use the workflow button (/) in the chat input to run a workflow.

​
Workflow Examples

​
Protein Structure Prediction Pipeline

Purpose: Complete structure analysis from sequence to visualization Steps:
  1. Validate the input protein sequence
  2. Search for similar structures in the PDB database
  3. Submit structure prediction using Boltz-2
  4. Wait for prediction to complete
  5. Visualize the predicted 3D structure
  6. Analyze prediction quality and write a report
Parameters:
  • SEQUENCE: Protein sequence to analyze
  • MODEL: Prediction tool (Boltz-2, Chai-1, or SimpleFold)

​
Literature Review Workflow

Purpose: Comprehensive literature search and summary Steps:
  1. Search PubMed for papers about TOPIC in last YEARS years
  2. Filter results by relevance and journal impact factor
  3. Retrieve the top COUNT most relevant papers
  4. Extract key findings from each paper
  5. Identify common themes and consensus
  6. List any conflicting results or debates
  7. Suggest potential research directions
  8. Generate a formatted bibliography
Parameters:
  • TOPIC: Research topic or question
  • YEARS: Time range to search (default: 5 years)
  • COUNT: Number of papers to analyze (default: 10)

​
Editing Workflows

1

Open Workflow

On the Workflows page, click on the workflow name to open it
2

Edit Mode

Click the “Edit” button in the top-right corner of the workflow detail modal
3

Make Changes

  • Modify workflow name or description
  • Add, remove, or reorder steps
  • Update parameter definitions
4

Save Changes

Click “Save” to update your workflow. You can also click “Cancel” to discard your changes.
You can also ask OpenBio to update the workflow for you.

​
Workflow Best Practices

Always create workflows from conversations that worked well. Test your steps manually before saving them as a workflow.
“Protein-Ligand Docking with Chai-1” is much better than “Workflow 1”. Include the key tools or methods in the name.
Explain what each step does, what inputs are needed, and what outputs to expect. Future you will thank you!
Include validation steps and error handling. What should happen if a file is missing or a search returns no results?
Run new workflows with test data first. Verify each step works as expected before using on important research.
Create smaller, focused workflows that can be combined rather than one giant workflow that tries to do everything.

​
Troubleshooting

Solutions:
  • Type / to see autocomplete suggestions
  • Check spelling of workflow name
  • Verify workflow is saved (check Workflows page)
Consider:
  • Some steps (like structure prediction) inherently take time
  • Check if prediction jobs are queued (normal during busy periods)
  • Break complex workflows into smaller, faster ones

​
Next Steps

Predictions

Use workflows with AI prediction tools

Chat Guide

Master chat to create better workflows

File Management

Organize files for workflow use

Projects

Organize workflows by project
Power User Tip: Create a “Workflow Development” project where you test and refine new workflows. Once they’re working perfectly, move them to your main research projects.