Bring the relevant context
Share the language, framework and relevant version when you know them. Include the code around the problem, the full error message and what you expected to happen. A small reproducible example is usually more useful than a screenshot of an unexplained failure.
For larger changes, explain the existing behavior that must remain unchanged. Ask for a focused fix before requesting a refactor. The model only has the context supplied to the conversation; do not assume it can see your entire repository or local environment.
This Python function returns the wrong result for an empty list. Explain the cause, propose the smallest fix, and add tests for an empty list, one item and repeated values.Try it in UnioAI
Work with files, not just pasted snippets
You can attach a file and ask UnioAI to explain or modify it, then download the resulting file. Describe the exact changes and any format requirements. For example, a text file containing Python code can be prepared as a .py file; changing the extension alone does not make arbitrary text executable.
Fast is useful for small code questions. Smart is intended for more complex analysis and changes involving several requirements. File size and account limits still apply. Keep credentials, private keys and production secrets out of uploaded material.
Check the result before running it
Review generated code and compare it with the original. Run your own tests in a suitable environment and inspect dependencies before installing them. UnioAI does not guarantee that a proposed fix compiles, passes your test suite or is secure.
When a question depends on a recent library version, web search can help find documentation. Check that the cited documentation matches the version your project actually uses. An answer for a newer release may not work in an older codebase.