Prompt builder
Fill in what you know: the task, the context, the format you want back. The builder assembles a clear, structured prompt that works across models. Copy it, download it or count its tokens.
The anatomy of a strong prompt
Most weak prompts fail for the same reason: the writer knows things the model doesn't. The builder walks you through the parts that close that gap.
- Task. One clear instruction with a verb: summarize, extract, rewrite, review, decide.
- Context. Who it's for, why it matters, what you'll do with the answer and what a great result looks like.
- Input. The material to work on, fenced off with tags so it can't be confused with instructions.
- Format. The exact shape you want back. For anything a program will read, ask for JSON with named fields.
- Rules and an example. Constraints the model can check itself against, and one good answer to imitate.
Then test it: run the prompt on five real inputs, note where it goes wrong, and add a rule or example for each failure. Our prompt engineering guide covers techniques, reasoning models and versioning.
Need hands-on help building it?
RAG, agents, fine-tuning or a production chatbot: describe the project and we'll suggest an approach and, if you want, partners who build it.
- Free and vendor-neutral: no model provider pays for placement.
- Your details go only to partners you agree to.
Frequently asked questions
Why use tags like <task> and <context> in a prompt?
Clear delimiters separate your instructions from the material the model should work on. That reduces confusion, makes long prompts easier to maintain, and lowers the risk that text inside a document is mistaken for an instruction. Most current models handle XML-style tags well.
Should I tell the model to think step by step?
For standard chat models, asking for step-by-step reasoning can improve answers to multi-step problems. Reasoning models already think before they answer, so providers generally advise against prescribing the steps; ask for the final answer and its format instead.
What makes a prompt good?
A specific task, the context a smart newcomer would need, an example or two of a good answer, and an explicit format. Most weak prompts fail because they leave out context the writer took for granted.