Max Tokens Guard

Enforce a maximum token limit on LLM requests.

Max Tokens Guard

This guard ensures that the max_tokens parameter sent to the LLM does not exceed the configured limit. If the parameter is absent, it adds the limit. If it’s present and exceeds the limit, it overwrites it.

Configuration

• max_tokens (int, optional): The maximum allowed tokens.

Configuration

Field	Type	Default	Description
max_tokens	`integer	null`	None

Examples

# Example 1
type: max_tokens
config:
  max_tokens: 1000