Interface JudgeAgentConfig

The criteria that the judge will use to evaluate the conversation.

Whether to pass audio content to the judge model.

• true / false — explicit; overrides auto-detection.
• null (default) — auto-detect: true when the conversation has audio AND the judge model is known to support multimodal input.

Set includeAudio: false as a cost-reduction escape hatch on multimodal models when audio evaluation is not needed.

Whether to include a structured voice timeline in the judge input.

• true / false — explicit.
• null (default) — auto: true when the conversation has audio.

Whether to include OTel / LangWatch trace spans in the judge input.

• true / false — explicit.
• null (default) — auto: true when LangWatch / OTel is configured.

Maximum number of tool-calling steps for progressive trace discovery. Only applies when the trace exceeds the token threshold.

Optional span collector for telemetry. Defaults to global singleton.

A custom system prompt to override the default behavior of the judge.

Token threshold for switching to structure-only trace rendering. When the full trace digest exceeds this estimated token count, the judge receives a structure-only view with expand_trace and grep_trace tools for progressive discovery.