Prompt Engineering Tools
Prompt Engineering covers the two things that decide whether an AI feature works and what it costs: the quality of the instructions and the number of tokens they take. The prompt builder assembles a role, task, context, examples and an output format into a prompt in Markdown or XML-tag style, and the quality checker scores an existing prompt and lists concrete fixes such as adding an output format or replacing vague words.
5 free tools · no sign-up · no ads
All Prompt Engineering (5)
AI Prompt Builder
Build structured prompts with role, context, examples and output formatPrompt Quality Checker
Score a prompt and get fixes for vague or missing instructionsLLM Token Counter
Count tokens for GPT, Claude, Gemini and Llama, with costPrompt Token Reducer
Cut tokens from prompts and context without losing meaningLLM API Cost Calculator
Monthly API cost by model, with caching and batch discounts
Which tool do I need?
Turn a rough request into a reliable prompt
Fill in the role, the task, the context and an example, and choose JSON or a table as the output.
Find out why a prompt gives inconsistent results
Get a score and a list of issues such as missing format, conflicting rules or placeholders left unfilled.
Check a document fits in the context window
Count its tokens and see the share of each model's window it takes.
Lower the token bill of a long system prompt
Strip markup and filler, minify JSON, then price the before and after per million calls.
About Prompt Engineering
The token tools deal with cost and limits. The token counter gives exact counts for OpenAI tokenizers and estimates for other model families, with the share of each context window used. The token reducer removes whitespace, markup and filler from prompts and pasted context, showing every change and the tokens saved. The cost calculator turns tokens per request and requests per day into a monthly bill, with prompt caching and batch discounts, and compares models side by side.
Guides
- How to Write Better AI Prompts: Structure, Examples and FormatThe parts of a prompt that change the answer: task, context, audience, examples, output format and rules phrased as what to do. With a before and after rewrite and the vendor advice behind each habit.
- How to Reduce LLM Token Usage and API CostWhere the tokens in an LLM bill come from, how to count them, and the levers that cut cost in order of impact: output length, prompt caching, batch APIs, model choice and trimming what you paste.
Frequently asked questions
Is prompt engineering still worth learning with newer models?
Yes, though it looks less like tricks and more like writing a clear brief. Newer models follow instructions well, so the gains come from giving context, an example and an explicit output format, which is what the builder and checker focus on.
Why does the same text have different token counts in different models?
Each model family uses its own tokenizer with its own vocabulary. Text in English usually lands around four characters per token, while code, numbers and other languages vary more, so counts for one model are only an estimate for another.
What is the quickest way to reduce LLM costs?
Shorten what you send on every call, such as long system prompts and pasted documents, cache the parts that repeat, and use a smaller model for simple steps. The cost calculator shows how much each of those changes saves.
Do output tokens cost more than input tokens?
For most hosted models, yes, often several times more per token. That is why asking for concise answers or structured output can cut a bill as much as trimming the prompt.