The Code Block allows you to run Python code securely during your flow’s execution using Azure Code Interpreter. This block provides a safe and isolated environment for executing custom Python code. Common use cases include:
  • Performing complex calculations
  • Implementing custom data processing logic
  • Using Python libraries and functions
  • Data transformation and analysis
The Code Block only supports Python code execution.
Code Block Screenshot

​
Inputs

​
input1
any
Default input port. Input names are configurable and each creates a corresponding port. Values are accessible in your Python code via the inputs dictionary with typed values (e.g., inputs["input1"]["type"], inputs["input1"]["value"]).

​
Outputs

​
output1
any
Default output port. Output names are configurable and each creates a corresponding port. The code must return a dictionary with typed values matching the configured output names (e.g., {"output1": {"type": "string", "value": "result"}}).

​
Editor Settings

​
Inputs
stringList
default:["input1"]
List of input names. Each name creates an input port on the block. Names are sanitized to be valid Python identifiers (only alphanumeric and underscore characters).
​
Outputs
stringList
default:["output1"]
List of output names. Each name creates an output port on the block. The code must return values for all configured outputs.
​
AI Assist
custom
An AI assistant to help write and modify your code.
​
Code
code
The Python code to execute. Must include a main(inputs) function that processes the inputs and returns a dictionary of outputs. Default template:
def main(inputs):
  return {
    "output1": {
      "type": inputs["input1"]["type"],
      "value": inputs["input1"]["value"]
    }
  }

​
Code Structure

Your code must:
  • Define a main(inputs) function
  • Access inputs through the inputs dictionary where each input has type and value properties
  • Return a dictionary with output names matching those configured in settings
  • Each output must have type and value properties