Create new prompt

JSON output

Ensure the output is valid JSON.

Code output

Ensure the output only contains code.

Profanity filter

Make sure the output doesn't contain any profanity.

Tool call

Ensure the output is a tool call.

PII redaction

Redact personally-identifiable information from the model response.

Specific topic

Ensure the output is related to a specific topic.

Output length

Limit the length of the output to a certain number of words.

Code execution

Prevent the model from generating or executing code.

Sentiment analysis

Analyze the sentiment of the output to ensure it is positive.

Hate speech detection

Detect and prevent hate speech in the output.

Factuality check

Verify the factual accuracy of the model output.