Skip to content

Generate and edit at the cursor

WhoseDev inline edit prompt in a Python file beside the AI chat panel

Use Ctrl+K on Windows/Linux or Cmd+K on macOS to open inline editing at the cursor or for a selected block of code.

Place the cursor where you want new code, open inline editing, and describe the result:

Add validation that rejects empty task titles.

Select code before opening inline editing. WhoseDev defaults to an edit action and lets you refine the selected code without losing the surrounding file context.

Use the action picker to choose:

  • Generate for code at the cursor.
  • Edit to rewrite a selection.
  • Explain or Quick Question to understand code without applying an edit.
  • Send to Chat to continue a larger task in the chat panel.

When inline editing is not enough — multi-file refactors, test runs, or terminal verification — run WhoseDev: Continue Inline Edit in Agent (whosedev.inlineEdit.continueInAgent) to hand the task to the chat agent with the inline context preserved.

Inline edit can fall back to a qualified local Ollama model when cloud routes fail:

Setting Default Purpose
whosedev.com.inlineEdit.localFallbackEnabled on Use local Ollama as final fallback for Generate/Edit when cloud fails or times out
whosedev.com.inlineEdit.allowSlowLocal off Allow CPU-slow local models; when off, slow models are skipped unless no cloud model is available

See Local AI and Offline Models for Ollama setup and qualification.

Read every suggested change before you keep it. If the result does not match your intent, undo it and make the request more specific, including constraints such as language version, error behavior, test coverage, or files that must not change.