A prompt is the input you send to a model: system instructions, user messages, and any context or examples. Prompt quality drives output quality, and prompt length drives cost. In Infere, prompts can be stored and versioned as PML files rather than hardcoded strings.

Why it matters

The prompt is the program. Small changes in wording, examples, or structure can move output quality more than swapping models - and longer prompts cost more on every request.

How it works

A prompt is usually a system message plus a sequence of user and assistant messages, and sometimes retrieved context or examples. Keeping prompts in version control rather than hardcoded strings makes them reviewable and reproducible.

Example

A prompt consisting of a system message, three worked examples, and the user's question is exactly what the model sees - and exactly what you are billed for.

← Back to the full glossary

Put the platform behind the terms

Route, evaluate, and monitor every AI request from one OpenAI-compatible platform.

Start Free → Explore the Features