Assistant prefill
An assistant prefill is text placed at the start of the model’s own reply so that it continues from there instead of choosing its own opening.
Prefilling with an opening brace forces valid JSON and removes the "Sure, here is the JSON you asked for" preamble. Prefilling with a code fence forces a code block. It is the cheapest way to control output shape, and it is available in the Anthropic API and several others.
Related terms
- System prompt
- A system prompt is a persistent instruction that sets a model’s role, rules, and boundaries for an entire conversation, separate from the individual messages a user sends.
- Structured output
- Structured output is model output constrained to a machine-readable shape such as JSON, CSV, or a fixed schema, rather than free prose.
More on core concepts
See the full glossary, read the guides, or put it into practice in the prompt builder.