fix(config): accept video and audio model inputs

Preserve configured audio/video model input modalities through provider catalog normalization.\n\nFixes #20721.\nThanks @alvinttang.

CHANGELOG.md

@@ -74,6 +74,8 @@ Docs: https://docs.openclaw.ai
 - Channels/status: keep read-only channel lists on manifest and package metadata by default, loading setup runtime only for explicit fallback callers. Thanks @shakkernerd.
 - Plugins/onboarding: defer onboarding install-record index writes until the guarded config commit so setup failures cannot leave the plugin index ahead of `openclaw.json`. Thanks @shakkernerd.
 - Plugins/registry: resolve web provider ownership from the installed plugin index instead of broad manifest scans on secret, tool, and pricing paths. Thanks @shakkernerd.
+- Config/providers: accept `video` and `audio` in configured model `input` values and
+  preserve them in provider catalog entries. Fixes #20721. Thanks @alvinttang.
 - TTS: strip model-emitted TTS directives from streamed block text before channel
   delivery, including directives split across adjacent blocks, while preserving
   the accumulated raw reply for final-mode synthesis. Fixes #38937.

extensions/lmstudio/src/models.ts

@@ -269,7 +269,7 @@ export type LmstudioModelBase = {
   trainedForToolUse: boolean;
   loaded: boolean;
   reasoning: boolean;
-  input: ModelDefinitionConfig["input"];
+  input: Array<"text" | "image">;
   cost: ModelDefinitionConfig["cost"];
   contextWindow: number;
   contextTokens: number;

extensions/lmstudio/src/setup.ts

@@ -822,6 +822,9 @@ export async function prepareLmstudioDynamicModels(
       provider: PROVIDER_ID,
       api: ctx.providerConfig?.api ?? `openai-completions`,
       baseUrl,
+      input: model.input.filter(
+        (entry): entry is "text" | "image" => entry === "text" || entry === "image",
+      ),
     }),
   );
 }

src/agents/model-catalog.types.ts

@@ -1,4 +1,4 @@
-export type ModelInputType = "text" | "image" | "document";
+export type ModelInputType = "text" | "image" | "audio" | "video" | "document";
 
 export type ModelCatalogEntry = {
   id: string;

src/config/schema.base.generated.ts

@@ -2908,6 +2908,14 @@ export const GENERATED_BASE_CONFIG_SCHEMA: BaseConfigSchemaResponse = {
                               type: "string",
                               const: "image",
                             },
+                            {
+                              type: "string",
+                              const: "video",
+                            },
+                            {
+                              type: "string",
+                              const: "audio",
+                            },
                           ],
                         },
                       },

src/config/types.models.ts

@@ -80,7 +80,7 @@ export type ModelDefinitionConfig = {
   api?: ModelApi;
   baseUrl?: string;
   reasoning: boolean;
-  input: Array<"text" | "image">;
+  input: Array<"text" | "image" | "video" | "audio">;
   cost: {
     input: number;
     output: number;

src/config/zod-schema.core.ts

@@ -312,7 +312,11 @@ export const ModelDefinitionSchema = z
     api: ModelApiSchema.optional(),
     baseUrl: z.string().min(1).optional(),
     reasoning: z.boolean().optional(),
-    input: z.array(z.union([z.literal("text"), z.literal("image")])).optional(),
+    input: z
+      .array(
+        z.union([z.literal("text"), z.literal("image"), z.literal("video"), z.literal("audio")]),
+      )
+      .optional(),
     cost: z
       .object({
         input: z.number().optional(),

src/plugin-sdk/lmstudio-runtime.ts

@@ -41,7 +41,7 @@ export type LmstudioModelBase = {
   trainedForToolUse: boolean;
   loaded: boolean;
   reasoning: boolean;
-  input: ModelDefinitionConfig["input"];
+  input: Array<"text" | "image">;
   cost: ModelDefinitionConfig["cost"];
   contextWindow: number;
   contextTokens: number;

src/plugin-sdk/provider-catalog-shared.test.ts

@@ -1,6 +1,7 @@
 import { describe, expect, it } from "vitest";
 import {
   applyProviderNativeStreamingUsageCompat,
+  readConfiguredProviderCatalogEntries,
   supportsNativeStreamingUsageCompat,
 } from "./provider-catalog-shared.js";
 import type { ModelDefinitionConfig } from "./provider-model-shared.js";
@@ -54,3 +55,43 @@ describe("provider-catalog-shared native streaming usage compat", () => {
     expect(provider.models?.[1]?.compat?.supportsUsageInStreaming).toBe(false);
   });
 });
+
+describe("provider-catalog-shared configured catalog entries", () => {
+  it("preserves configured audio and video input modalities", () => {
+    expect(
+      readConfiguredProviderCatalogEntries({
+        providerId: "kilocode",
+        config: {
+          models: {
+            providers: {
+              kilocode: {
+                baseUrl: "https://api.kilo.ai/api/gateway/",
+                api: "openai-completions",
+                models: [
+                  {
+                    id: "google/gemini-3-pro-preview",
+                    name: "Gemini 3 Pro Preview",
+                    input: ["text", "image", "video", "audio"],
+                    reasoning: true,
+                    cost: { input: 0, output: 0, cacheRead: 0, cacheWrite: 0 },
+                    contextWindow: 1048576,
+                    maxTokens: 65536,
+                  },
+                ],
+              },
+            },
+          },
+        },
+      }),
+    ).toEqual([
+      {
+        provider: "kilocode",
+        id: "google/gemini-3-pro-preview",
+        name: "Gemini 3 Pro Preview",
+        input: ["text", "image", "video", "audio"],
+        reasoning: true,
+        contextWindow: 1048576,
+      },
+    ]);
+  });
+});

src/plugin-sdk/provider-catalog-shared.ts

@@ -23,7 +23,7 @@ export type ConfiguredProviderCatalogEntry = {
   provider: string;
   contextWindow?: number;
   reasoning?: boolean;
-  input?: Array<"text" | "image" | "document">;
+  input?: Array<"text" | "image" | "audio" | "video" | "document">;
 };
 
 function normalizeConfiguredCatalogModelInput(
@@ -33,8 +33,12 @@ function normalizeConfiguredCatalogModelInput(
     return undefined;
   }
   const normalized = input.filter(
-    (item): item is "text" | "image" | "document" =>
+    (item): item is "text" | "image" | "audio" | "video" | "document" =>
-      item === "text" || item === "image" || item === "document",
+      item === "text" ||
+      item === "image" ||
+      item === "audio" ||
+      item === "video" ||
+      item === "document",
   );
   return normalized.length > 0 ? normalized : undefined;
 }