fix(minimax): transcode voice-note tts to opus

CHANGELOG.md

@@ -91,7 +91,7 @@ Docs: https://docs.openclaw.ai
 - Plugins/OpenCode: strip unsupported disabled Responses reasoning payloads for OpenCode image understanding. Fixes #70252.
 - Plugins/OpenCode/OpenCode Go: register image understanding metadata so the image tool is available for OpenCode catalog models with vision support. Fixes #70482 and #61789.
 - Providers/ElevenLabs: omit the MP3-only `Accept` header for PCM telephony synthesis, so Voice Call requests for `pcm_22050` no longer receive MP3 audio. Fixes #67340. Thanks @marcchabot.
-- Providers/MiniMax TTS: mark MP3 output voice-compatible for Telegram voice-note delivery. Fixes #63540.
+- Providers/MiniMax TTS: transcode voice-note targets to Opus so Feishu/Telegram receive native voice messages instead of MP3 file attachments. Fixes #63540, #64134, and #70445.
 - Providers/Microsoft TTS: keep allowlisted bundled speech providers discoverable even when another speech plugin has already registered, so Edge/Microsoft TTS is available alongside OpenAI. Fixes #62117 and #66850.
 - Providers/Microsoft TTS: honor legacy `messages.tts.providers.edge` voice settings after normalizing Edge TTS to the Microsoft provider. Fixes #64153.
 - Providers/OpenRouter: add an OpenRouter TTS provider using the OpenAI-compatible `/audio/speech` endpoint and `OPENROUTER_API_KEY`. Fixes #71268.

docs/providers/minimax.md

@@ -244,6 +244,18 @@ exposed separately through the plugin-owned `MiniMax-VL-01` media provider.
 See [Image Generation](/tools/image-generation) for shared tool parameters, provider selection, and failover behavior.
 </Note>
 
+### Text-to-speech
+
+The bundled `minimax` plugin registers MiniMax T2A v2 as a speech provider for
+`messages.tts`.
+
+- Default TTS model: `speech-2.8-hd`
+- Default voice: `English_expressive_narrator`
+- Normal audio attachments stay MP3.
+- Voice-note targets such as Feishu and Telegram are transcoded from MiniMax
+  MP3 to 48kHz Opus with `ffmpeg`, because the Feishu/Lark file API only
+  accepts `file_type: "opus"` for native audio messages.
+
 ### Music generation
 
 The bundled `minimax` plugin also registers music generation through the shared

docs/tools/tts.md

@@ -488,7 +488,7 @@ These override `messages.tts.*` for that host.
   - 48kHz / 64kbps is a good voice message tradeoff.
 - **Other channels**: MP3 (`mp3_44100_128` from ElevenLabs, `mp3` from OpenAI).
   - 44.1kHz / 128kbps is the default balance for speech clarity.
-- **MiniMax**: MP3 (`speech-2.8-hd` model, 32kHz sample rate). Voice-note format not natively supported; use OpenAI or ElevenLabs for guaranteed Opus voice messages.
+- **MiniMax**: MP3 (`speech-2.8-hd` model, 32kHz sample rate) for normal audio attachments. For voice-note targets such as Feishu and Telegram, OpenClaw transcodes the MiniMax MP3 to 48kHz Opus with `ffmpeg` before delivery.
 - **Google Gemini**: Gemini API TTS returns raw 24kHz PCM. OpenClaw wraps it as WAV for audio attachments and returns PCM directly for Talk/telephony. Native Opus voice-note format is not supported by this path.
 - **Gradium**: WAV for audio attachments, Opus for voice-note targets, and `ulaw_8000` at 8 kHz for telephony.
 - **xAI**: MP3 by default; `responseFormat` may be `mp3`, `wav`, `pcm`, `mulaw`, or `alaw`. OpenClaw uses xAI's batch REST TTS endpoint and returns a complete audio attachment; xAI's streaming TTS WebSocket is not used by this provider path. Native Opus voice-note format is not supported by this path.

extensions/minimax/speech-provider.test.ts

@@ -1,4 +1,11 @@
 import { afterEach, beforeEach, describe, expect, it, vi } from "vitest";
+
+const runFfmpegMock = vi.hoisted(() => vi.fn());
+
+vi.mock("openclaw/plugin-sdk/media-runtime", () => ({
+  runFfmpeg: runFfmpegMock,
+}));
+
 import { buildMinimaxSpeechProvider } from "./speech-provider.js";
 
 describe("buildMinimaxSpeechProvider", () => {
@@ -213,6 +220,7 @@ describe("buildMinimaxSpeechProvider", () => {
 
     beforeEach(() => {
       vi.stubGlobal("fetch", vi.fn());
+      runFfmpegMock.mockReset();
     });
 
     afterEach(() => {
@@ -240,7 +248,7 @@ describe("buildMinimaxSpeechProvider", () => {
 
       expect(result.outputFormat).toBe("mp3");
       expect(result.fileExtension).toBe(".mp3");
-      expect(result.voiceCompatible).toBe(true);
+      expect(result.voiceCompatible).toBe(false);
       expect(result.audioBuffer.toString()).toBe("fake-audio-data");
 
       expect(mockFetch).toHaveBeenCalledOnce();
@@ -250,6 +258,44 @@ describe("buildMinimaxSpeechProvider", () => {
       expect(body.model).toBe("speech-2.8-hd");
       expect(body.text).toBe("Hello world");
       expect(body.voice_setting.voice_id).toBe("English_expressive_narrator");
+      expect(runFfmpegMock).not.toHaveBeenCalled();
+    });
+
+    it("transcodes MiniMax MP3 to Opus for voice-note targets", async () => {
+      const hexAudio = Buffer.from("fake-mp3-data").toString("hex");
+      const mockFetch = vi.mocked(globalThis.fetch);
+      mockFetch.mockResolvedValueOnce(
+        new Response(JSON.stringify({ data: { audio: hexAudio } }), {
+          status: 200,
+          headers: { "Content-Type": "application/json" },
+        }),
+      );
+      runFfmpegMock.mockImplementationOnce(async (args: string[]) => {
+        const outputPath = args.at(-1);
+        if (typeof outputPath !== "string") {
+          throw new Error("missing ffmpeg output path");
+        }
+        await import("node:fs/promises").then((fs) =>
+          fs.writeFile(outputPath, Buffer.from("fake-opus-data")),
+        );
+      });
+
+      const result = await provider.synthesize({
+        text: "Hello world",
+        cfg: {} as never,
+        providerConfig: { apiKey: "sk-test", baseUrl: "https://api.minimaxi.com" },
+        target: "voice-note",
+        timeoutMs: 30000,
+      });
+
+      expect(result.outputFormat).toBe("opus");
+      expect(result.fileExtension).toBe(".opus");
+      expect(result.voiceCompatible).toBe(true);
+      expect(result.audioBuffer.toString()).toBe("fake-opus-data");
+      expect(runFfmpegMock).toHaveBeenCalledWith(
+        expect.arrayContaining(["-c:a", "libopus", "-ar", "48000"]),
+        { timeoutMs: 30000 },
+      );
     });
 
     it("applies overrides", async () => {

extensions/minimax/speech-provider.ts

@@ -1,3 +1,6 @@
+import { mkdir, mkdtemp, readFile, rm, writeFile } from "node:fs/promises";
+import path from "node:path";
+import { runFfmpeg } from "openclaw/plugin-sdk/media-runtime";
 import { normalizeResolvedSecretInputString } from "openclaw/plugin-sdk/secret-input";
 import type {
   SpeechDirectiveTokenParseContext,
@@ -6,6 +9,7 @@ import type {
   SpeechProviderPlugin,
 } from "openclaw/plugin-sdk/speech-core";
 import { asFiniteNumber, asObject, trimToUndefined } from "openclaw/plugin-sdk/speech-core";
+import { resolvePreferredOpenClawTmpDir } from "openclaw/plugin-sdk/temp-path";
 import {
   DEFAULT_MINIMAX_TTS_BASE_URL,
   MINIMAX_TTS_MODELS,
@@ -150,6 +154,41 @@ function parseDirectiveToken(ctx: SpeechDirectiveTokenParseContext): {
   }
 }
 
+async function transcodeMp3ToOpus(audioBuffer: Buffer, timeoutMs: number | undefined) {
+  const tempRoot = resolvePreferredOpenClawTmpDir();
+  await mkdir(tempRoot, { recursive: true, mode: 0o700 });
+  const tempDir = await mkdtemp(path.join(tempRoot, "tts-minimax-"));
+  try {
+    const inputPath = path.join(tempDir, "input.mp3");
+    const outputPath = path.join(tempDir, "voice.opus");
+    await writeFile(inputPath, audioBuffer, { mode: 0o600 });
+    await runFfmpeg(
+      [
+        "-hide_banner",
+        "-loglevel",
+        "error",
+        "-y",
+        "-i",
+        inputPath,
+        "-vn",
+        "-c:a",
+        "libopus",
+        "-b:a",
+        "64k",
+        "-ar",
+        "48000",
+        "-ac",
+        "1",
+        outputPath,
+      ],
+      { timeoutMs },
+    );
+    return await readFile(outputPath);
+  } finally {
+    await rm(tempDir, { recursive: true, force: true });
+  }
+}
+
 export function buildMinimaxSpeechProvider(): SpeechProviderPlugin {
   return {
     id: "minimax",
@@ -223,11 +262,20 @@ export function buildMinimaxSpeechProvider(): SpeechProviderPlugin {
         pitch: overrides.pitch ?? config.pitch,
         timeoutMs: req.timeoutMs,
       });
+      if (req.target === "voice-note") {
+        const opusBuffer = await transcodeMp3ToOpus(audioBuffer, req.timeoutMs);
+        return {
+          audioBuffer: opusBuffer,
+          outputFormat: "opus",
+          fileExtension: ".opus",
+          voiceCompatible: true,
+        };
+      }
       return {
         audioBuffer,
         outputFormat: "mp3",
         fileExtension: ".mp3",
-        voiceCompatible: true,
+        voiceCompatible: false,
       };
     },
   };