fix(models): support minimax-portal coding plan vlm routing for image tool (openclaw#33953)

Verified:
- pnpm install --frozen-lockfile
- pnpm build
- pnpm check
- pnpm test:macmini

Co-authored-by: tars90percent <252094836+tars90percent@users.noreply.github.com>

CHANGELOG.md

@@ -134,6 +134,7 @@ Docs: https://docs.openclaw.ai
 - Routing/legacy route guard tightening: require legacy session-key channel hints to match the saved delivery channel before inheriting external routing metadata, preventing custom namespaced keys like `agent:<agent>:work:<ticket>` from inheriting stale non-webchat routes.
 - Gateway/internal client routing continuity: prevent webchat/TUI/UI turns from inheriting stale external reply routes by requiring explicit `deliver: true` for external delivery, keeping main-session external inheritance scoped to non-Webchat/UI clients, and honoring configured `session.mainKey` when identifying main-session continuity. (from #35321, #34635, #35356) Thanks @alexyyyander and @Octane0411.
 - Security/auth labels: remove token and API-key snippets from user-facing auth status labels so `/status` and `/models` do not expose credential fragments. (#33262) thanks @cu1ch3n.
+- Models/MiniMax portal vision routing: add `MiniMax-VL-01` to the `minimax-portal` provider, route portal image understanding through the MiniMax VLM endpoint, and align media auto-selection plus Telegram sticker description with the shared portal image provider path. (#33953) Thanks @tars90percent.
 - Auth/credential semantics: align profile eligibility + probe diagnostics with SecretRef/expiry rules and harden browser download atomic writes. (#33733) thanks @joshavant.
 - Security/audit denyCommands guidance: suggest likely exact node command IDs for unknown `gateway.nodes.denyCommands` entries so ineffective denylist entries are easier to correct. (#29713) thanks @liquidhorizon88-bot.
 - Agents/overload failover handling: classify overloaded provider failures separately from rate limits/status timeouts, add short overload backoff before retry/failover, record overloaded prompt/assistant failures as transient auth-profile cooldowns (with probeable same-provider fallback) instead of treating them like persistent auth/billing failures, and keep one-shot cron retry classification aligned so overloaded fallback summaries still count as transient retries.

src/agents/minimax-vlm.normalizes-api-key.test.ts

@@ -45,3 +45,14 @@ describe("minimaxUnderstandImage apiKey normalization", () => {
     await runNormalizationCase("minimax-\u0417\u2502test-key");
   });
 });
+
+describe("isMinimaxVlmModel", () => {
+  it("only matches the canonical MiniMax VLM model id", async () => {
+    const { isMinimaxVlmModel } = await import("./minimax-vlm.js");
+
+    expect(isMinimaxVlmModel("minimax", "MiniMax-VL-01")).toBe(true);
+    expect(isMinimaxVlmModel("minimax-portal", "MiniMax-VL-01")).toBe(true);
+    expect(isMinimaxVlmModel("minimax-portal", "custom-vision")).toBe(false);
+    expect(isMinimaxVlmModel("openai", "MiniMax-VL-01")).toBe(false);
+  });
+});

src/agents/minimax-vlm.ts

@@ -6,6 +6,14 @@ type MinimaxBaseResp = {
   status_msg?: string;
 };
 
+export function isMinimaxVlmProvider(provider: string): boolean {
+  return provider === "minimax" || provider === "minimax-portal";
+}
+
+export function isMinimaxVlmModel(provider: string, modelId: string): boolean {
+  return isMinimaxVlmProvider(provider) && modelId.trim() === "MiniMax-VL-01";
+}
+
 function coerceApiHost(params: {
   apiHost?: string;
   modelBaseUrl?: string;

src/agents/models-config.providers.nvidia.test.ts

@@ -71,10 +71,9 @@ describe("MiniMax implicit provider (#15275)", () => {
             "minimax-portal:default": {
               type: "oauth",
               provider: "minimax-portal",
-              oauth: {
+              access: "token",
-                access: "token",
+              refresh: "refresh-token",
-                expires: Date.now() + 60_000,
+              expires: Date.now() + 60_000,
-              },
             },
           },
         },
@@ -87,6 +86,18 @@ describe("MiniMax implicit provider (#15275)", () => {
     const providers = await resolveImplicitProviders({ agentDir });
     expect(providers?.["minimax-portal"]?.authHeader).toBe(true);
   });
+
+  it("should include minimax portal provider when MINIMAX_OAUTH_TOKEN is configured", async () => {
+    const agentDir = mkdtempSync(join(tmpdir(), "openclaw-test-"));
+    await withEnvAsync({ MINIMAX_OAUTH_TOKEN: "portal-token" }, async () => {
+      const providers = await resolveImplicitProviders({ agentDir });
+      expect(providers?.["minimax-portal"]).toBeDefined();
+      expect(providers?.["minimax-portal"]?.authHeader).toBe(true);
+      expect(providers?.["minimax-portal"]?.models?.some((m) => m.id === "MiniMax-VL-01")).toBe(
+        true,
+      );
+    });
+  });
 });
 
 describe("vLLM provider", () => {

src/agents/models-config.providers.ts

@@ -771,6 +771,12 @@ function buildMinimaxPortalProvider(): ProviderConfig {
     api: "anthropic-messages",
     authHeader: true,
     models: [
+      buildMinimaxModel({
+        id: MINIMAX_DEFAULT_VISION_MODEL_ID,
+        name: "MiniMax VL 01",
+        reasoning: false,
+        input: ["text", "image"],
+      }),
       buildMinimaxTextModel({
         id: MINIMAX_DEFAULT_MODEL_ID,
         name: "MiniMax M2.5",
@@ -1116,8 +1122,9 @@ export async function resolveImplicitProviders(params: {
     providers.minimax = { ...buildMinimaxProvider(), apiKey: minimaxKey };
   }
 
+  const minimaxPortalEnvKey = resolveEnvApiKeyVarName("minimax-portal");
   const minimaxOauthProfile = listProfilesForProvider(authStore, "minimax-portal");
-  if (minimaxOauthProfile.length > 0) {
+  if (minimaxPortalEnvKey || minimaxOauthProfile.length > 0) {
     providers["minimax-portal"] = {
       ...buildMinimaxPortalProvider(),
       apiKey: MINIMAX_OAUTH_MARKER,

src/agents/tools/image-tool.test.ts

@@ -273,6 +273,32 @@ describe("image tool implicit imageModel config", () => {
     });
   });
 
+  it("pairs minimax-portal primary with MiniMax-VL-01 (and fallbacks) when auth exists", async () => {
+    await withTempAgentDir(async (agentDir) => {
+      await writeAuthProfiles(agentDir, {
+        version: 1,
+        profiles: {
+          "minimax-portal:default": {
+            type: "oauth",
+            provider: "minimax-portal",
+            access: "oauth-test",
+            refresh: "refresh-test",
+            expires: Date.now() + 60_000,
+          },
+        },
+      });
+      vi.stubEnv("OPENAI_API_KEY", "openai-test");
+      vi.stubEnv("ANTHROPIC_API_KEY", "anthropic-test");
+      const cfg: OpenClawConfig = {
+        agents: { defaults: { model: { primary: "minimax-portal/MiniMax-M2.5" } } },
+      };
+      expect(resolveImageModelConfigForTool({ cfg, agentDir })).toEqual(
+        createDefaultImageFallbackExpectation("minimax-portal/MiniMax-VL-01"),
+      );
+      expect(createImageTool({ config: cfg, agentDir })).not.toBeNull();
+    });
+  });
+
   it("pairs zai primary with glm-4.6v (and fallbacks) when auth exists", async () => {
     await withTempAgentDir(async (agentDir) => {
       vi.stubEnv("ZAI_API_KEY", "zai-test");

src/agents/tools/image-tool.ts

@@ -3,7 +3,7 @@ import { Type } from "@sinclair/typebox";
 import type { OpenClawConfig } from "../../config/config.js";
 import { resolveUserPath } from "../../utils.js";
 import { loadWebMedia } from "../../web/media.js";
-import { minimaxUnderstandImage } from "../minimax-vlm.js";
+import { isMinimaxVlmModel, isMinimaxVlmProvider, minimaxUnderstandImage } from "../minimax-vlm.js";
 import {
   coerceImageAssistantText,
   coerceImageModelConfig,
@@ -110,8 +110,8 @@ export function resolveImageModelConfigForTool(params: {
   let preferred: string | null = null;
 
   // MiniMax users: always try the canonical vision model first when auth exists.
-  if (primary.provider === "minimax" && providerOk) {
+  if (isMinimaxVlmProvider(primary.provider) && providerOk) {
-    preferred = "minimax/MiniMax-VL-01";
+    preferred = `${primary.provider}/MiniMax-VL-01`;
   } else if (providerOk && providerVisionFromConfig) {
     preferred = providerVisionFromConfig;
   } else if (primary.provider === "zai" && providerOk) {
@@ -229,7 +229,7 @@ async function runImagePrompt(params: {
       });
 
       // MiniMax VLM only supports a single image; use the first one.
-      if (model.provider === "minimax") {
+      if (isMinimaxVlmModel(model.provider, model.id)) {
         const first = params.images[0];
         const imageDataUrl = `data:${first.mimeType};base64,${first.base64}`;
         const text = await minimaxUnderstandImage({

src/media-understanding/defaults.test.ts

@@ -1,8 +1,10 @@
 import { describe, expect, it } from "vitest";
 import {
   AUTO_AUDIO_KEY_PROVIDERS,
+  AUTO_IMAGE_KEY_PROVIDERS,
   AUTO_VIDEO_KEY_PROVIDERS,
   DEFAULT_AUDIO_MODELS,
+  DEFAULT_IMAGE_MODELS,
 } from "./defaults.js";
 
 describe("DEFAULT_AUDIO_MODELS", () => {
@@ -22,3 +24,15 @@ describe("AUTO_VIDEO_KEY_PROVIDERS", () => {
     expect(AUTO_VIDEO_KEY_PROVIDERS).toContain("moonshot");
   });
 });
+
+describe("AUTO_IMAGE_KEY_PROVIDERS", () => {
+  it("includes minimax-portal auto key resolution", () => {
+    expect(AUTO_IMAGE_KEY_PROVIDERS).toContain("minimax-portal");
+  });
+});
+
+describe("DEFAULT_IMAGE_MODELS", () => {
+  it("includes the MiniMax portal vision default", () => {
+    expect(DEFAULT_IMAGE_MODELS["minimax-portal"]).toBe("MiniMax-VL-01");
+  });
+});

src/media-understanding/defaults.ts

@@ -46,6 +46,7 @@ export const AUTO_IMAGE_KEY_PROVIDERS = [
   "anthropic",
   "google",
   "minimax",
+  "minimax-portal",
   "zai",
 ] as const;
 export const AUTO_VIDEO_KEY_PROVIDERS = ["google", "moonshot"] as const;
@@ -54,6 +55,7 @@ export const DEFAULT_IMAGE_MODELS: Record<string, string> = {
   anthropic: "claude-opus-4-6",
   google: "gemini-3-flash-preview",
   minimax: "MiniMax-VL-01",
+  "minimax-portal": "MiniMax-VL-01",
   zai: "glm-4.6v",
 };
 export const CLI_OUTPUT_MAX_BUFFER = 5 * MB;

src/media-understanding/providers/image.test.ts

@@ -0,0 +1,133 @@
+import { beforeEach, describe, expect, it, vi } from "vitest";
+
+const completeMock = vi.fn();
+const minimaxUnderstandImageMock = vi.fn();
+const ensureOpenClawModelsJsonMock = vi.fn(async () => {});
+const getApiKeyForModelMock = vi.fn(async () => ({
+  apiKey: "oauth-test",
+  source: "test",
+  mode: "oauth",
+}));
+const requireApiKeyMock = vi.fn((auth: { apiKey?: string }) => auth.apiKey ?? "");
+const setRuntimeApiKeyMock = vi.fn();
+const discoverModelsMock = vi.fn();
+
+vi.mock("@mariozechner/pi-ai", async (importOriginal) => {
+  const actual = await importOriginal<typeof import("@mariozechner/pi-ai")>();
+  return {
+    ...actual,
+    complete: completeMock,
+  };
+});
+
+vi.mock("../../agents/minimax-vlm.js", () => ({
+  isMinimaxVlmProvider: (provider: string) =>
+    provider === "minimax" || provider === "minimax-portal",
+  isMinimaxVlmModel: (provider: string, modelId: string) =>
+    (provider === "minimax" || provider === "minimax-portal") && modelId === "MiniMax-VL-01",
+  minimaxUnderstandImage: minimaxUnderstandImageMock,
+}));
+
+vi.mock("../../agents/models-config.js", () => ({
+  ensureOpenClawModelsJson: ensureOpenClawModelsJsonMock,
+}));
+
+vi.mock("../../agents/model-auth.js", () => ({
+  getApiKeyForModel: getApiKeyForModelMock,
+  requireApiKey: requireApiKeyMock,
+}));
+
+vi.mock("../../agents/pi-model-discovery-runtime.js", () => ({
+  discoverAuthStorage: () => ({
+    setRuntimeApiKey: setRuntimeApiKeyMock,
+  }),
+  discoverModels: discoverModelsMock,
+}));
+
+describe("describeImageWithModel", () => {
+  beforeEach(() => {
+    vi.clearAllMocks();
+    minimaxUnderstandImageMock.mockResolvedValue("portal ok");
+    discoverModelsMock.mockReturnValue({
+      find: vi.fn(() => ({
+        provider: "minimax-portal",
+        id: "MiniMax-VL-01",
+        input: ["text", "image"],
+        baseUrl: "https://api.minimax.io/anthropic",
+      })),
+    });
+  });
+
+  it("routes minimax-portal image models through the MiniMax VLM endpoint", async () => {
+    const { describeImageWithModel } = await import("./image.js");
+
+    const result = await describeImageWithModel({
+      cfg: {},
+      agentDir: "/tmp/openclaw-agent",
+      provider: "minimax-portal",
+      model: "MiniMax-VL-01",
+      buffer: Buffer.from("png-bytes"),
+      fileName: "image.png",
+      mime: "image/png",
+      prompt: "Describe the image.",
+      timeoutMs: 1000,
+    });
+
+    expect(result).toEqual({
+      text: "portal ok",
+      model: "MiniMax-VL-01",
+    });
+    expect(ensureOpenClawModelsJsonMock).toHaveBeenCalled();
+    expect(getApiKeyForModelMock).toHaveBeenCalled();
+    expect(requireApiKeyMock).toHaveBeenCalled();
+    expect(setRuntimeApiKeyMock).toHaveBeenCalledWith("minimax-portal", "oauth-test");
+    expect(minimaxUnderstandImageMock).toHaveBeenCalledWith({
+      apiKey: "oauth-test",
+      prompt: "Describe the image.",
+      imageDataUrl: `data:image/png;base64,${Buffer.from("png-bytes").toString("base64")}`,
+      modelBaseUrl: "https://api.minimax.io/anthropic",
+    });
+    expect(completeMock).not.toHaveBeenCalled();
+  });
+
+  it("uses generic completion for non-canonical minimax-portal image models", async () => {
+    discoverModelsMock.mockReturnValue({
+      find: vi.fn(() => ({
+        provider: "minimax-portal",
+        id: "custom-vision",
+        input: ["text", "image"],
+        baseUrl: "https://api.minimax.io/anthropic",
+      })),
+    });
+    completeMock.mockResolvedValue({
+      role: "assistant",
+      api: "anthropic-messages",
+      provider: "minimax-portal",
+      model: "custom-vision",
+      stopReason: "stop",
+      timestamp: Date.now(),
+      content: [{ type: "text", text: "generic ok" }],
+    });
+
+    const { describeImageWithModel } = await import("./image.js");
+
+    const result = await describeImageWithModel({
+      cfg: {},
+      agentDir: "/tmp/openclaw-agent",
+      provider: "minimax-portal",
+      model: "custom-vision",
+      buffer: Buffer.from("png-bytes"),
+      fileName: "image.png",
+      mime: "image/png",
+      prompt: "Describe the image.",
+      timeoutMs: 1000,
+    });
+
+    expect(result).toEqual({
+      text: "generic ok",
+      model: "custom-vision",
+    });
+    expect(completeMock).toHaveBeenCalledOnce();
+    expect(minimaxUnderstandImageMock).not.toHaveBeenCalled();
+  });
+});

src/media-understanding/providers/image.ts

@@ -1,6 +1,6 @@
 import type { Api, Context, Model } from "@mariozechner/pi-ai";
 import { complete } from "@mariozechner/pi-ai";
-import { minimaxUnderstandImage } from "../../agents/minimax-vlm.js";
+import { isMinimaxVlmModel, minimaxUnderstandImage } from "../../agents/minimax-vlm.js";
 import { getApiKeyForModel, requireApiKey } from "../../agents/model-auth.js";
 import { ensureOpenClawModelsJson } from "../../agents/models-config.js";
 import { coerceImageAssistantText } from "../../agents/tools/image-tool.helpers.js";
@@ -40,7 +40,7 @@ export async function describeImageWithModel(
   authStorage.setRuntimeApiKey(model.provider, apiKey);
 
   const base64 = params.buffer.toString("base64");
-  if (model.provider === "minimax") {
+  if (isMinimaxVlmModel(model.provider, model.id)) {
     const text = await minimaxUnderstandImage({
       apiKey,
       prompt: params.prompt ?? "Describe the image.",

src/media-understanding/providers/index.test.ts

@@ -24,4 +24,12 @@ describe("media-understanding provider registry", () => {
     expect(provider?.id).toBe("moonshot");
     expect(provider?.capabilities).toEqual(["image", "video"]);
   });
+
+  it("registers the minimax portal provider", () => {
+    const registry = buildMediaUnderstandingRegistry();
+    const provider = getMediaUnderstandingProvider("minimax-portal", registry);
+
+    expect(provider?.id).toBe("minimax-portal");
+    expect(provider?.capabilities).toEqual(["image"]);
+  });
 });

src/media-understanding/providers/index.ts

@@ -4,7 +4,7 @@ import { anthropicProvider } from "./anthropic/index.js";
 import { deepgramProvider } from "./deepgram/index.js";
 import { googleProvider } from "./google/index.js";
 import { groqProvider } from "./groq/index.js";
-import { minimaxProvider } from "./minimax/index.js";
+import { minimaxPortalProvider, minimaxProvider } from "./minimax/index.js";
 import { mistralProvider } from "./mistral/index.js";
 import { moonshotProvider } from "./moonshot/index.js";
 import { openaiProvider } from "./openai/index.js";
@@ -16,6 +16,7 @@ const PROVIDERS: MediaUnderstandingProvider[] = [
   googleProvider,
   anthropicProvider,
   minimaxProvider,
+  minimaxPortalProvider,
   moonshotProvider,
   mistralProvider,
   zaiProvider,

src/media-understanding/providers/minimax/index.ts

@@ -6,3 +6,9 @@ export const minimaxProvider: MediaUnderstandingProvider = {
   capabilities: ["image"],
   describeImage: describeImageWithModel,
 };
+
+export const minimaxPortalProvider: MediaUnderstandingProvider = {
+  id: "minimax-portal",
+  capabilities: ["image"],
+  describeImage: describeImageWithModel,
+};

src/telegram/sticker-cache.ts

@@ -12,6 +12,7 @@ import type { OpenClawConfig } from "../config/config.js";
 import { STATE_DIR } from "../config/paths.js";
 import { logVerbose } from "../globals.js";
 import { loadJsonFile, saveJsonFile } from "../infra/json-file.js";
+import { AUTO_IMAGE_KEY_PROVIDERS, DEFAULT_IMAGE_MODELS } from "../media-understanding/defaults.js";
 import { resolveAutoImageModel } from "../media-understanding/runner.js";
 
 const CACHE_FILE = path.join(STATE_DIR, "telegram", "sticker-cache.json");
@@ -142,7 +143,6 @@ export function getCacheStats(): { count: number; oldestAt?: string; newestAt?:
 
 const STICKER_DESCRIPTION_PROMPT =
   "Describe this sticker image in 1-2 sentences. Focus on what the sticker depicts (character, object, action, emotion). Be concise and objective.";
-const VISION_PROVIDERS = ["openai", "anthropic", "google", "minimax"] as const;
 let imageRuntimePromise: Promise<
   typeof import("../media-understanding/providers/image-runtime.js")
 > | null = null;
@@ -198,14 +198,7 @@ export async function describeStickerImage(params: DescribeStickerParams): Promi
     if (entries.length === 0) {
       return undefined;
     }
-    const defaultId =
+    const defaultId = DEFAULT_IMAGE_MODELS[provider];
-      provider === "openai"
-        ? "gpt-5-mini"
-        : provider === "anthropic"
-          ? "claude-opus-4-6"
-          : provider === "google"
-            ? "gemini-3-flash-preview"
-            : "MiniMax-VL-01";
     const preferred = entries.find((entry) => entry.id === defaultId);
     return preferred ?? entries[0];
   };
@@ -213,14 +206,16 @@ export async function describeStickerImage(params: DescribeStickerParams): Promi
   let resolved = null as { provider: string; model?: string } | null;
   if (
     activeModel &&
-    VISION_PROVIDERS.includes(activeModel.provider as (typeof VISION_PROVIDERS)[number]) &&
+    AUTO_IMAGE_KEY_PROVIDERS.includes(
+      activeModel.provider as (typeof AUTO_IMAGE_KEY_PROVIDERS)[number],
+    ) &&
     (await hasProviderKey(activeModel.provider))
   ) {
     resolved = activeModel;
   }
 
   if (!resolved) {
-    for (const provider of VISION_PROVIDERS) {
+    for (const provider of AUTO_IMAGE_KEY_PROVIDERS) {
       if (!(await hasProviderKey(provider))) {
         continue;
       }