AxBalancerOptions

type AxBalancerOptions<TModelKey> = object;

Defined in: https://github.com/ax-llm/ax/blob/e690b4823b8309e258594064abe941d4cff6ef89/src/ax/ai/balance.ts#L87

Options for the balancer.

Type Parameters

Properties

comparator()?

optional comparator: (a: AxAIService<unknown, unknown, TModelKey>, b: AxAIService<unknown, unknown, TModelKey>) => number;

Defined in: https://github.com/ax-llm/ax/blob/e690b4823b8309e258594064abe941d4cff6ef89/src/ax/ai/balance.ts#L88

Parameters

Returns

number

debug?

optional debug: boolean;

Defined in: https://github.com/ax-llm/ax/blob/e690b4823b8309e258594064abe941d4cff6ef89/src/ax/ai/balance.ts#L92

initialBackoffMs?

optional initialBackoffMs: number;

Defined in: https://github.com/ax-llm/ax/blob/e690b4823b8309e258594064abe941d4cff6ef89/src/ax/ai/balance.ts#L93

maxBackoffMs?

optional maxBackoffMs: number;

Defined in: https://github.com/ax-llm/ax/blob/e690b4823b8309e258594064abe941d4cff6ef89/src/ax/ai/balance.ts#L94

maxRetries?

optional maxRetries: number;

Defined in: https://github.com/ax-llm/ax/blob/e690b4823b8309e258594064abe941d4cff6ef89/src/ax/ai/balance.ts#L95

strategy?

optional strategy: AxBalancerAdaptiveStrategy<TModelKey>;

Defined in: https://github.com/ax-llm/ax/blob/e690b4823b8309e258594064abe941d4cff6ef89/src/ax/ai/balance.ts#L101

Opt into request-time provider selection using learned failure rate, successful latency, and estimated cost. When omitted, the existing comparator-ordered behavior is unchanged.