AI assistance
Get past the line
that has you stuck.
Ask about an unfamiliar function, work through an error or explore another approach. Use the answer as a starting point you can test in your own code.
A better question gets you closer
Give the assistant something specific to work with.
Understand
Make the code readable.
Ask for an explanation that matches what you are learning. A small example is often more useful than a long definition.
“Explain this function line by line. Show the result for a list containing 2, 4 and 6.”
Investigate
Work through an error.
Describe what you expected, what happened and the error you saw. That gives the assistant a concrete problem to examine.
“I expected a total of 30, but this returns 12. Which part of my loop should I check?”
Improve
Find a clearer approach.
Once your script works, ask how to make a small part easier to follow. Keep the version you understand while you compare.
“Can you simplify this function without changing its result? Explain each change.”
On iPhone & iPad
Keep the question close to your script.
The mobile Code Assistant can help generate, refine and debug Python inside the editor.
Try a suggestion, run the script and check the output. You decide which changes belong in your code.
Assistant features may require a subscription and your own provider API key.
A useful way to ask
1. Say what you want the script to do.
2. Include the smallest relevant piece of code.
3. Describe the result or error you see.
4. Ask for a change you can test.
AI in the Mac workspace
Ask with your project in view.
The Mac assistant can work with selected file or project context. Choose a supported provider, including local Ollama models.
For a multi-file project, give the assistant the context it needs for the question. Keep control of what you share and review the suggested changes.
Explore the Mac workspace →Before you use an assistant
Know what you are connecting.
Is AI access included in the app price?
An app subscription and an external AI service are separate. Depending on the provider, you may need an account, an API key or paid usage.
Does every answer work?
No. AI suggestions can be wrong. Read the change, run the script and check the result before relying on it.
Can I use a local model on Mac?
The direct Mac workspace supports configured Ollama models. Running a local model requires Ollama and a suitable model on your Mac.
What should I share?
Include the code needed to answer the question. Leave out passwords, API keys and private data you do not want to send to the provider.
Keep learning while you build.
Explore the editor, examples and tools around your next Python question.