diff --git a/astrbot/core/config/default.py b/astrbot/core/config/default.py
index eac4bafc01..ccf57127fd 100644
--- a/astrbot/core/config/default.py
+++ b/astrbot/core/config/default.py
@@ -1770,6 +1770,10 @@
                         "api_base": "https://api.minimax.chat/v1/t2a_v2",
                         "minimax-group-id": "",
                         "model": "speech-02-turbo",
+                        "minimax-voice-clone-audio": "",
+                        "minimax-voice-clone-id": "",
+                        "minimax-voice-clone-model": "speech-2.8-hd",
+                        "minimax-voice-clone-api-base": "https://api.minimaxi.com/v1",
                         "minimax-langboost": "auto",
                         "minimax-voice-speed": 1.0,
                         "minimax-voice-vol": 1.0,
@@ -2566,6 +2570,36 @@
                         "description": "单一音色",
                         "hint": "单一音色编号, 详见官网文档",
                     },
+                    "minimax-voice-clone-audio": {
+                        "type": "string",
+                        "description": "Voice clone audio path",
+                        "hint": "Local .mp3, .m4a, or .wav reference audio. Leave empty to disable automatic voice cloning.",
+                    },
+                    "minimax-voice-clone-id": {
+                        "type": "string",
+                        "description": "Voice clone ID",
+                        "hint": "Custom voice ID to create before synthesis. Required when voice clone audio is configured.",
+                    },
+                    "minimax-voice-clone-model": {
+                        "type": "string",
+                        "description": "Voice clone model",
+                        "options": [
+                            "speech-2.8-hd",
+                            "speech-2.6-hd",
+                            "speech-02-hd",
+                            "speech-01-hd",
+                        ],
+                        "hint": "HD speech model used to create the custom voice.",
+                    },
+                    "minimax-voice-clone-api-base": {
+                        "type": "string",
+                        "description": "Voice clone API base",
+                        "options": [
+                            "https://api.minimax.io/v1",
+                            "https://api.minimaxi.com/v1",
+                        ],
+                        "hint": "Regional MiniMax API base for file upload and voice cloning.",
+                    },
                     "minimax-voice-emotion": {
                         "type": "string",
                         "description": "情绪",
diff --git a/astrbot/core/provider/sources/minimax_tts_api_source.py b/astrbot/core/provider/sources/minimax_tts_api_source.py
index 97d746c557..309bbee782 100644
--- a/astrbot/core/provider/sources/minimax_tts_api_source.py
+++ b/astrbot/core/provider/sources/minimax_tts_api_source.py
@@ -1,7 +1,9 @@
+import asyncio
 import json
 import os
 import uuid
 from collections.abc import AsyncIterator
+from pathlib import Path
 
 import aiohttp
 
@@ -32,6 +34,20 @@ def __init__(
         )
         self.group_id: str = provider_config.get("minimax-group-id", "")
         self.set_model(provider_config.get("model", ""))
+        self.voice_clone_audio: str = str(
+            provider_config.get("minimax-voice-clone-audio") or "",
+        ).strip()
+        self.voice_clone_id: str = str(
+            provider_config.get("minimax-voice-clone-id") or "",
+        ).strip()
+        self.voice_clone_model: str = str(
+            provider_config.get("minimax-voice-clone-model") or "",
+        ).strip()
+        self.voice_clone_api_base: str = str(
+            provider_config.get("minimax-voice-clone-api-base") or "",
+        ).strip()
+        self._voice_clone_ready = False
+        self._voice_clone_lock = asyncio.Lock()
         self.lang_boost: str = provider_config.get("minimax-langboost", "auto")
         self.is_timber_weight: bool = provider_config.get(
             "minimax-is-timber-weight",
@@ -84,6 +100,155 @@ def __init__(
             "content-type": "application/json",
         }
 
+    def _voice_clone_url(self, path: str) -> str:
+        """Build a MiniMax voice-cloning endpoint URL."""
+        if self.voice_clone_api_base:
+            api_base = self.voice_clone_api_base
+        else:
+            api_base = self.api_base.rstrip("/")
+            if not api_base.endswith("/v1"):
+                api_base = api_base.rsplit("/", 1)[0]
+        return f"{api_base.rstrip('/')}/{path.lstrip('/')}"
+
+    async def _ensure_voice_clone(self) -> None:
+        """Create the configured custom voice before the first synthesis request."""
+        if not self.voice_clone_audio:
+            return
+        if self._voice_clone_ready:
+            return
+
+        async with self._voice_clone_lock:
+            if self._voice_clone_ready:
+                return
+            if not self.voice_clone_id:
+                raise ValueError(
+                    "MiniMax voice cloning requires 'minimax-voice-clone-id'.",
+                )
+            if not self.voice_clone_model:
+                raise ValueError(
+                    "MiniMax voice cloning requires 'minimax-voice-clone-model'.",
+                )
+            if not self.voice_clone_model.endswith("-hd"):
+                raise ValueError(
+                    "MiniMax voice cloning requires an HD speech model.",
+                )
+            if self.is_timber_weight:
+                raise ValueError(
+                    "MiniMax voice cloning cannot be combined with mixed voices.",
+                )
+
+            audio_path = Path(self.voice_clone_audio).expanduser()
+            if not audio_path.is_file():
+                raise FileNotFoundError(
+                    f"MiniMax voice-clone audio file does not exist: {audio_path}",
+                )
+
+            content_type = {
+                ".m4a": "audio/mp4",
+                ".mp3": "audio/mpeg",
+                ".wav": "audio/wav",
+            }.get(audio_path.suffix.lower())
+            if content_type is None:
+                raise ValueError(
+                    "MiniMax voice cloning supports only .mp3, .m4a, and .wav files.",
+                )
+
+            try:
+                async with aiohttp.ClientSession() as session:
+                    form = aiohttp.FormData()
+                    with audio_path.open("rb") as audio_file:
+                        form.add_field(
+                            "file",
+                            audio_file,
+                            filename=audio_path.name,
+                            content_type=content_type,
+                        )
+                        form.add_field("purpose", "voice_clone")
+                        async with session.post(
+                            self._voice_clone_url("files/upload"),
+                            headers={"Authorization": self.headers["Authorization"]},
+                            data=form,
+                            timeout=aiohttp.ClientTimeout(total=60),
+                        ) as response:
+                            if response.status >= 400:
+                                error_text = (await response.text())[:1024]
+                                raise RuntimeError(
+                                    "MiniMax voice-clone audio upload failed: "
+                                    f"HTTP {response.status}: {error_text}",
+                                )
+                            upload_data = await response.json(content_type=None)
+
+                    upload_base_resp = upload_data.get("base_resp", {})
+                    upload_status_code = upload_base_resp.get("status_code")
+                    if upload_status_code not in (None, 0, "0"):
+                        status_msg = upload_base_resp.get(
+                            "status_msg",
+                            "unknown error",
+                        )
+                        raise RuntimeError(
+                            f"MiniMax voice-clone audio upload failed: {status_msg}",
+                        )
+
+                    file_data = upload_data.get("file") or {}
+                    nested_data = upload_data.get("data") or {}
+                    file_id = (
+                        upload_data.get("file_id")
+                        or file_data.get("file_id")
+                        or nested_data.get("file_id")
+                    )
+                    if not file_id:
+                        raise RuntimeError(
+                            "MiniMax voice-clone audio upload returned no file_id.",
+                        )
+
+                    async with session.post(
+                        self._voice_clone_url("voice_clone"),
+                        headers=self.headers,
+                        json={
+                            "file_id": file_id,
+                            "voice_id": self.voice_clone_id,
+                            "model": self.voice_clone_model,
+                        },
+                        timeout=aiohttp.ClientTimeout(total=60),
+                    ) as response:
+                        if response.status >= 400:
+                            error_text = (await response.text())[:1024]
+                            raise RuntimeError(
+                                "MiniMax voice cloning failed: "
+                                f"HTTP {response.status}: {error_text}",
+                            )
+                        clone_data = await response.json(content_type=None)
+
+                base_resp = clone_data.get("base_resp", {})
+                status_code = base_resp.get("status_code")
+                if status_code not in (None, 0, "0"):
+                    status_msg = base_resp.get("status_msg", "unknown error")
+                    raise RuntimeError(
+                        f"MiniMax voice cloning failed: {status_msg}",
+                    )
+
+                clone_result = clone_data.get("data") or {}
+                voice_id = (
+                    clone_data.get("voice_id")
+                    or clone_result.get("voice_id")
+                    or self.voice_clone_id
+                )
+                self.voice_setting["voice_id"] = voice_id
+                self._voice_clone_ready = True
+            except aiohttp.ClientError as exc:
+                raise RuntimeError(
+                    f"MiniMax voice cloning request failed: {exc!s}",
+                ) from exc
+
+    async def clone_voice(self) -> str:
+        """Create and return the configured MiniMax custom voice ID."""
+        await self._ensure_voice_clone()
+        if not self.voice_clone_audio:
+            raise ValueError(
+                "MiniMax voice cloning requires 'minimax-voice-clone-audio'.",
+            )
+        return self.voice_setting["voice_id"]
+
     def _build_tts_stream_body(self, text: str):
         """构建流式请求体"""
         dict_body: dict[str, object] = {
@@ -154,6 +319,7 @@ async def _audio_play(self, audio_stream: AsyncIterator[str]) -> bytes:
         return b"".join(chunks)
 
     async def get_audio(self, text: str) -> str:
+        await self._ensure_voice_clone()
         temp_dir = get_astrbot_temp_path()
         os.makedirs(temp_dir, exist_ok=True)
         path = os.path.join(temp_dir, f"minimax_tts_api_{uuid.uuid4()}.wav")
diff --git a/dashboard/src/i18n/locales/en-US/features/config-metadata.json b/dashboard/src/i18n/locales/en-US/features/config-metadata.json
index fdd430c63e..72b27eafc9 100644
--- a/dashboard/src/i18n/locales/en-US/features/config-metadata.json
+++ b/dashboard/src/i18n/locales/en-US/features/config-metadata.json
@@ -1564,6 +1564,22 @@
         "description": "Single voice",
         "hint": "Single voice ID; see the official documentation."
       },
+      "minimax-voice-clone-audio": {
+        "description": "Voice clone audio path",
+        "hint": "Local .mp3, .m4a, or .wav reference audio. Leave empty to disable automatic voice cloning."
+      },
+      "minimax-voice-clone-id": {
+        "description": "Voice clone ID",
+        "hint": "Custom voice ID to create before synthesis. Required when voice clone audio is configured."
+      },
+      "minimax-voice-clone-model": {
+        "description": "Voice clone model",
+        "hint": "HD speech model used to create the custom voice."
+      },
+      "minimax-voice-clone-api-base": {
+        "description": "Voice clone API base",
+        "hint": "Regional MiniMax API base for file upload and voice cloning."
+      },
       "minimax-voice-emotion": {
         "description": "Emotion",
         "hint": "Controls emotion of synthesized speech. When set to auto, it selects emotion based on text."
diff --git a/dashboard/src/i18n/locales/ru-RU/features/config-metadata.json b/dashboard/src/i18n/locales/ru-RU/features/config-metadata.json
index 3cd0104ca1..fca8d0b852 100644
--- a/dashboard/src/i18n/locales/ru-RU/features/config-metadata.json
+++ b/dashboard/src/i18n/locales/ru-RU/features/config-metadata.json
@@ -1561,6 +1561,22 @@
                 "description": "Одиночный голос",
                 "hint": "ID одиночного голоса; см. официальную документацию."
             },
+            "minimax-voice-clone-audio": {
+                "description": "Voice clone audio path",
+                "hint": "Local .mp3, .m4a, or .wav reference audio. Leave empty to disable automatic voice cloning."
+            },
+            "minimax-voice-clone-id": {
+                "description": "Voice clone ID",
+                "hint": "Custom voice ID to create before synthesis. Required when voice clone audio is configured."
+            },
+            "minimax-voice-clone-model": {
+                "description": "Voice clone model",
+                "hint": "HD speech model used to create the custom voice."
+            },
+            "minimax-voice-clone-api-base": {
+                "description": "Voice clone API base",
+                "hint": "Regional MiniMax API base for file upload and voice cloning."
+            },
             "minimax-voice-emotion": {
                 "description": "Эмоция",
                 "hint": "Эмоция речи. 'auto' выбирает эмоцию на основе текста."
diff --git a/dashboard/src/i18n/locales/zh-CN/features/config-metadata.json b/dashboard/src/i18n/locales/zh-CN/features/config-metadata.json
index e25cb8e0fb..46058f596f 100644
--- a/dashboard/src/i18n/locales/zh-CN/features/config-metadata.json
+++ b/dashboard/src/i18n/locales/zh-CN/features/config-metadata.json
@@ -1566,6 +1566,22 @@
         "description": "单一音色",
         "hint": "单一音色编号, 详见官网文档"
       },
+      "minimax-voice-clone-audio": {
+        "description": "Voice clone audio path",
+        "hint": "Local .mp3, .m4a, or .wav reference audio. Leave empty to disable automatic voice cloning."
+      },
+      "minimax-voice-clone-id": {
+        "description": "Voice clone ID",
+        "hint": "Custom voice ID to create before synthesis. Required when voice clone audio is configured."
+      },
+      "minimax-voice-clone-model": {
+        "description": "Voice clone model",
+        "hint": "HD speech model used to create the custom voice."
+      },
+      "minimax-voice-clone-api-base": {
+        "description": "Voice clone API base",
+        "hint": "Regional MiniMax API base for file upload and voice cloning."
+      },
       "minimax-voice-emotion": {
         "description": "情绪",
         "hint": "控制合成语音的情绪。当为 auto 时，将根据文本内容自动选择情绪。"
diff --git a/tests/test_minimax_tts_api_source.py b/tests/test_minimax_tts_api_source.py
new file mode 100644
index 0000000000..aa66d53782
--- /dev/null
+++ b/tests/test_minimax_tts_api_source.py
@@ -0,0 +1,205 @@
+import asyncio
+import json
+from pathlib import Path
+
+import pytest
+
+from astrbot.core.provider.sources.minimax_tts_api_source import (
+    ProviderMiniMaxTTSAPI,
+)
+
+
+class _FakeResponse:
+    def __init__(self, payload: dict, status: int = 200) -> None:
+        self.payload = payload
+        self.status = status
+
+    async def __aenter__(self):
+        return self
+
+    async def __aexit__(self, *_args) -> None:
+        return None
+
+    async def json(self, **_kwargs) -> dict:
+        return self.payload
+
+    async def text(self) -> str:
+        return json.dumps(self.payload)
+
+
+class _FakeSession:
+    def __init__(self, responses: list[_FakeResponse]) -> None:
+        self.responses = responses
+        self.posts: list[tuple[str, dict]] = []
+
+    async def __aenter__(self):
+        return self
+
+    async def __aexit__(self, *_args) -> None:
+        return None
+
+    def post(self, url: str, **kwargs):
+        self.posts.append((url, kwargs))
+        return self.responses.pop(0)
+
+
+class _FakeFormData:
+    last: "_FakeFormData | None" = None
+
+    def __init__(self) -> None:
+        self.fields: list[tuple[str, object, dict]] = []
+        self.__class__.last = self
+
+    def add_field(self, name: str, value: object, **kwargs) -> None:
+        self.fields.append((name, value, kwargs))
+
+
+def _make_provider(audio_path: Path, **overrides) -> ProviderMiniMaxTTSAPI:
+    config = {
+        "id": "test-minimax-tts",
+        "type": "minimax_tts_api",
+        "api_key": "test-key",
+        "api_base": "https://api.minimaxi.com/v1/t2a_v2",
+        "model": "speech-02-turbo",
+        "minimax-voice-id": "default_voice",
+        "minimax-voice-clone-audio": str(audio_path),
+        "minimax-voice-clone-id": "custom_voice",
+        "minimax-voice-clone-model": "speech-2.8-hd",
+        "minimax-voice-clone-api-base": "https://api.minimaxi.com/v1",
+    }
+    config.update(overrides)
+    return ProviderMiniMaxTTSAPI(config, {})
+
+
+@pytest.mark.asyncio
+async def test_clone_voice_uploads_audio_and_uses_returned_voice_id(
+    monkeypatch,
+    tmp_path,
+):
+    audio_path = tmp_path / "reference.wav"
+    audio_path.write_bytes(b"wav data")
+    provider = _make_provider(audio_path)
+    session = _FakeSession(
+        [
+            _FakeResponse({"file": {"file_id": "file-123"}}),
+            _FakeResponse({"data": {"voice_id": "created-voice"}}),
+        ]
+    )
+    monkeypatch.setattr(
+        "astrbot.core.provider.sources.minimax_tts_api_source.aiohttp.ClientSession",
+        lambda: session,
+    )
+    monkeypatch.setattr(
+        "astrbot.core.provider.sources.minimax_tts_api_source.aiohttp.FormData",
+        _FakeFormData,
+    )
+
+    assert await provider.clone_voice() == "created-voice"
+    assert provider.voice_setting["voice_id"] == "created-voice"
+    assert (
+        json.loads(provider._build_tts_stream_body("hello"))["voice_setting"][
+            "voice_id"
+        ]
+        == "created-voice"
+    )
+    assert len(session.posts) == 2
+    assert session.posts[0][0] == "https://api.minimaxi.com/v1/files/upload"
+    assert session.posts[0][1]["headers"] == {"Authorization": "Bearer test-key"}
+    assert session.posts[1][0] == "https://api.minimaxi.com/v1/voice_clone"
+    assert session.posts[1][1]["json"] == {
+        "file_id": "file-123",
+        "voice_id": "custom_voice",
+        "model": "speech-2.8-hd",
+    }
+    assert _FakeFormData.last is not None
+    assert [field[0] for field in _FakeFormData.last.fields] == ["file", "purpose"]
+    assert _FakeFormData.last.fields[1][1] == "voice_clone"
+
+
+@pytest.mark.asyncio
+@pytest.mark.parametrize(
+    "api_base",
+    ["https://api.minimax.io/v1", "https://api.minimaxi.com/v1"],
+)
+async def test_clone_voice_supports_both_regional_api_bases(
+    monkeypatch,
+    tmp_path,
+    api_base,
+):
+    audio_path = tmp_path / "reference.mp3"
+    audio_path.write_bytes(b"mp3 data")
+    provider = _make_provider(
+        audio_path,
+        **{"minimax-voice-clone-api-base": api_base},
+    )
+    session = _FakeSession([_FakeResponse({"file_id": "file-123"}), _FakeResponse({})])
+    monkeypatch.setattr(
+        "astrbot.core.provider.sources.minimax_tts_api_source.aiohttp.ClientSession",
+        lambda: session,
+    )
+    monkeypatch.setattr(
+        "astrbot.core.provider.sources.minimax_tts_api_source.aiohttp.FormData",
+        _FakeFormData,
+    )
+
+    assert await provider.clone_voice() == "custom_voice"
+    assert [post[0] for post in session.posts] == [
+        f"{api_base}/files/upload",
+        f"{api_base}/voice_clone",
+    ]
+
+
+@pytest.mark.asyncio
+async def test_clone_voice_is_single_flight(monkeypatch, tmp_path):
+    audio_path = tmp_path / "reference.wav"
+    audio_path.write_bytes(b"wav data")
+    provider = _make_provider(audio_path)
+    session = _FakeSession([_FakeResponse({"file_id": "file-123"}), _FakeResponse({})])
+    monkeypatch.setattr(
+        "astrbot.core.provider.sources.minimax_tts_api_source.aiohttp.ClientSession",
+        lambda: session,
+    )
+    monkeypatch.setattr(
+        "astrbot.core.provider.sources.minimax_tts_api_source.aiohttp.FormData",
+        _FakeFormData,
+    )
+
+    assert await asyncio.gather(provider.clone_voice(), provider.clone_voice()) == [
+        "custom_voice",
+        "custom_voice",
+    ]
+    assert len(session.posts) == 2
+
+
+@pytest.mark.asyncio
+async def test_clone_voice_validates_configuration_and_audio_format(tmp_path):
+    audio_path = tmp_path / "reference.txt"
+    audio_path.write_text("not audio")
+    provider = _make_provider(audio_path, **{"minimax-voice-clone-id": ""})
+    with pytest.raises(ValueError, match="minimax-voice-clone-id"):
+        await provider.clone_voice()
+
+    provider = _make_provider(audio_path)
+    with pytest.raises(ValueError, match="only \\.mp3, \\.m4a, and \\.wav"):
+        await provider.clone_voice()
+
+
+@pytest.mark.asyncio
+async def test_clone_voice_propagates_api_status_error(monkeypatch, tmp_path):
+    audio_path = tmp_path / "reference.wav"
+    audio_path.write_bytes(b"wav data")
+    provider = _make_provider(audio_path)
+    session = _FakeSession(
+        [_FakeResponse({"base_resp": {"status_code": 1004, "status_msg": "bad audio"}})]
+    )
+    monkeypatch.setattr(
+        "astrbot.core.provider.sources.minimax_tts_api_source.aiohttp.ClientSession",
+        lambda: session,
+    )
+    monkeypatch.setattr(
+        "astrbot.core.provider.sources.minimax_tts_api_source.aiohttp.FormData",
+        _FakeFormData,
+    )
+
+    with pytest.raises(RuntimeError, match="bad audio"):
+        await provider.clone_voice()