While the model is working, the cell shows #GETTING_DATA. This is a waiting status. It changes to the answer automatically, and you can keep using Excel.
PROMPT returns one finished answer in one cell. Use it for summaries, translations, classifications or explanations.
=XLLAMA.PROMPT("Summarize this feedback", A2:B5)Syntax: =XLLAMA.PROMPT(prompt, context, [model])
- prompt: what you want the model to do.
- context: the cells the model should read.
- model: optional model name, such as
"gemma3:4b". Omit it to use your selected model.
PROMPT answers are plain text, without markdown or an introduction. They use temperature 0 and a fixed seed, so the same inputs give consistent results. If something goes wrong, the cell shows a plain-language message, such as Ollama isn’t running, model isn’t downloaded or Excel was busy.
PROMPT reuses a saved answer when the inputs, model and relevant settings match. It does not request another answer on every unchanged recalculation. For fresh answers, select the XLlama formula cells and choose Settings → Refresh selected formula answers. XLlama regenerates those results; other cells keep their displayed answers. Refresh up to 1,000 formula cells at a time. Cells with identical inputs share the refreshed cache when they next recalculate. Clear formula cache remains available to clear all saved formula answers.
Free supports up to five XLlama formula cells across the workbook. New formula answers share the 20-response monthly allowance with chat; cached answers do not use another response. Refreshing answers or clearing the cache can generate new responses. Pro removes these XLlama limits.
After updating to XLlama 1.2.0, existing XLlama formula cells regenerate once because the formula prompt changed. On Free, each regeneration uses one of the month’s responses.
Use commas between arguments, or semicolons if your Excel uses them. The optional final model argument overrides the selected model. Select at most 1,000 input cells; large tasks work best in smaller batches. Review AI results before relying on them.