Skip to main content
In a few minutes, you’ll have a working automation that fills out a web form, extracts structured data, and runs deterministically in production.

​
Create your first workflow

1

Create a new workflow

Click Create new workflow from your Dashboard. We provide the starting URL and test data the agent will need to complete the task below.We’ll use a U.S. Department of Transportation form with VIN (vehicle identification number) data:
Starting URL
https://vpic.nhtsa.dot.gov/decoder
Example Test Data
{
  "VIN": "1G1ZE5ST8HF198765",
  "Year": "2017"
}
2

Prompt the agent

Prompt the agent through chat to start the task. The agent generates and executes code to follow your instructions.We’ll start with a simple prompt.
Example prompt
Enter the requested information and decode VIN.
You can continue prompting to build on top of the workflow, including scraping information from the page.In this example, we’ll scrape all the vehicle information found.
Scrape vehicle info
Scroll up and scrape all vehicle information from the page.
3

View Output

The scraper agent will create a JavaScript script and corresponding JSON file with the following output.
workspace/scraper_outputs/vehicle_info.json
{
  "vin": "1G1ZE5ST8HF198765",
  "totalAttributes": 140,
  "attributes": [
    {
      "groupName": "",
      "element": "Suggested VIN",
      "value": ""
    },
    {
      "groupName": "",
      "element": "Error Code",
      "value": "1"
    },
    {
      "groupName": "",
      "element": "Error Text",
      "value": "1 - Check Digit (9th position) does not calculate properly"
    },
    {
      "groupName": "",
      "element": "Vehicle Descriptor",
      "value": "1G1ZE5ST*HF"
    },
    {
      "groupName": "General",
      "element": "Make",
      "value": "CHEVROLET"
    },
    {
      "groupName": "General",
      "element": "Manufacturer Name",
      "value": "GENERAL MOTORS LLC"
    },
    {
      "groupName": "General",
      "element": "Model",
      "value": "Malibu"
    },
    {
      "groupName": "General",
      "element": "Model Year",
      "value": "2017"
    },
    {
      "groupName": "General",
      "element": "Plant City",
      "value": "KANSAS CITY"
    },
    {
      "groupName": "General",
      "element": "Trim",
      "value": "LT (1LT)"
    },
    {
      "groupName": "General",
      "element": "Vehicle Type",
      "value": "PASSENGER CAR"
    },
    {
      "groupName": "General",
      "element": "Plant Country",
      "value": "UNITED STATES (USA)"
    }
  ]
}
This is a simple example to get you started. In production, Simplex handles forms with 100+ fields, branching logic, and multi-step flows across complex web portals.

​
Run the workflow

Now that we’ve built our workflow, let’s run it in production.
1

Step 1: Open production view

Click Workflow Runs in the top right for your newly created workflow, then click Run Production Workflow.
2

Step 2: Fill variables

Simplex automatically generates a variables schema for our production workflow based on the shape of the test data you provide.With the example case, this means that the production workflow expects a VIN and Year field, both strings.We’ll test our production flow with the same data for verification and just do 1 run.
Example Test Data
{
  "VIN": "1G1ZE5ST8HF198765",
  "Year": "2017"
}
3

Step 3: Run the workflow

Click Run workflow to execute the workflow.
You can also trigger the same workflow programmatically via the API or SDK: 
from simplex import SimplexClient

client = SimplexClient(api_key="your-api-key")
result = client.run_workflow("your-workflow-id", variables={"VIN": "1G1ZE5ST8HF198765", "Year": "2017"})
print(f"Session started: {result['session_id']}")

​
Next steps