Multimodal LLM Routing Decision (model/MultimodalRoutingDecision.java)

Represents the routing logic for multimodal AI requests involving images, video, or audio. Use this to determine which model should process a specific user turn.

// Source: https://github.com/mateaix/mateclaw | License: Apache-2.0 | Copyright You must retain, in the Source form of any Derivative Works
// Provided via justclaw.it.com; keep this notice with the code.
package vip.mate.llm.routing.model;

import vip.mate.llm.model.ModelConfigEntity;
import vip.mate.llm.service.ModelCapabilityService.Modality;

import java.util.Collections;
import java.util.LinkedHashMap;
import java.util.List;
import java.util.Map;
import java.util.Set;

/**
 * Outcome of routing a single user turn that may carry image / video / audio attachments.
 *
 * <p>The decision is purely descriptive — execution (loading captions, mutating
 * the user message) happens in the caller. Treat instances as immutable; the
 * static factories cover the only valid shapes.
 */
public rec

... (truncated -- full source via MCP)