Merge pull request #4299 from open-webui/dev-elevenlabs

feat: Fetch ElevenLabs voice ID by name

backend/apps/audio/main.py

@@ -10,12 +10,12 @@ from fastapi import (
     File,
     Form,
 )
-
 from fastapi.responses import StreamingResponse, JSONResponse, FileResponse
 
 from fastapi.middleware.cors import CORSMiddleware
 from pydantic import BaseModel
 
+from typing import List
 import uuid
 import requests
 import hashlib
@@ -31,6 +31,7 @@ from utils.utils import (
 )
 from utils.misc import calculate_sha256
 
+
 from config import (
     SRC_LOG_LEVELS,
     CACHE_DIR,
@@ -252,15 +253,15 @@ async def speech(request: Request, user=Depends(get_verified_user)):
             )
 
     elif app.state.config.TTS_ENGINE == "elevenlabs":
-
         payload = None
         try:
             payload = json.loads(body.decode("utf-8"))
         except Exception as e:
             log.exception(e)
-            pass
+            raise HTTPException(status_code=400, detail="Invalid JSON payload")
 
-        url = f"https://api.elevenlabs.io/v1/text-to-speech/{payload['voice']}"
+        voice_id = payload.get("voice", "")
+        url = f"https://api.elevenlabs.io/v1/text-to-speech/{voice_id}"
 
         headers = {
             "Accept": "audio/mpeg",
@@ -435,3 +436,69 @@ def transcribe(
             status_code=status.HTTP_400_BAD_REQUEST,
             detail=ERROR_MESSAGES.DEFAULT(e),
         )
+
+
+def get_available_models() -> List[dict]:
+    if app.state.config.TTS_ENGINE == "openai":
+        return [{"id": "tts-1"}, {"id": "tts-1-hd"}]
+    elif app.state.config.TTS_ENGINE == "elevenlabs":
+        headers = {
+            "xi-api-key": app.state.config.TTS_API_KEY,
+            "Content-Type": "application/json",
+        }
+
+        try:
+            response = requests.get(
+                "https://api.elevenlabs.io/v1/models", headers=headers
+            )
+            response.raise_for_status()
+            models = response.json()
+            return [
+                {"name": model["name"], "id": model["model_id"]} for model in models
+            ]
+        except requests.RequestException as e:
+            log.error(f"Error fetching voices: {str(e)}")
+    return []
+
+
+@app.get("/models")
+async def get_models(user=Depends(get_verified_user)):
+    return {"models": get_available_models()}
+
+
+def get_available_voices() -> List[dict]:
+    if app.state.config.TTS_ENGINE == "openai":
+        return [
+            {"name": "alloy", "id": "alloy"},
+            {"name": "echo", "id": "echo"},
+            {"name": "fable", "id": "fable"},
+            {"name": "onyx", "id": "onyx"},
+            {"name": "nova", "id": "nova"},
+            {"name": "shimmer", "id": "shimmer"},
+        ]
+    elif app.state.config.TTS_ENGINE == "elevenlabs":
+        headers = {
+            "xi-api-key": app.state.config.TTS_API_KEY,
+            "Content-Type": "application/json",
+        }
+
+        try:
+            response = requests.get(
+                "https://api.elevenlabs.io/v1/voices", headers=headers
+            )
+            response.raise_for_status()
+            voices_data = response.json()
+
+            voices = []
+            for voice in voices_data.get("voices", []):
+                voices.append({"name": voice["name"], "id": voice["voice_id"]})
+            return voices
+        except requests.RequestException as e:
+            log.error(f"Error fetching voices: {str(e)}")
+
+    return []
+
+
+@app.get("/voices")
+async def get_voices(user=Depends(get_verified_user)):
+    return {"voices": get_available_voices()}

src/lib/apis/audio/index.ts

@@ -131,3 +131,59 @@ export const synthesizeOpenAISpeech = async (
 
 	return res;
 };
+
+export const getModels = async (token: string = '') => {
+	let error = null;
+
+	const res = await fetch(`${AUDIO_API_BASE_URL}/models`, {
+		method: 'GET',
+		headers: {
+			'Content-Type': 'application/json',
+			Authorization: `Bearer ${token}`
+		}
+	})
+		.then(async (res) => {
+			if (!res.ok) throw await res.json();
+			return res.json();
+		})
+		.catch((err) => {
+			error = err.detail;
+			console.log(err);
+
+			return null;
+		});
+
+	if (error) {
+		throw error;
+	}
+
+	return res;
+};
+
+export const getVoices = async (token: string = '') => {
+	let error = null;
+
+	const res = await fetch(`${AUDIO_API_BASE_URL}/voices`, {
+		method: 'GET',
+		headers: {
+			'Content-Type': 'application/json',
+			Authorization: `Bearer ${token}`
+		}
+	})
+		.then(async (res) => {
+			if (!res.ok) throw await res.json();
+			return res.json();
+		})
+		.catch((err) => {
+			error = err.detail;
+			console.log(err);
+
+			return null;
+		});
+
+	if (error) {
+		throw error;
+	}
+
+	return res;
+};

src/lib/components/admin/Settings/Audio.svelte

 <script lang="ts">
-	import { getAudioConfig, updateAudioConfig } from '$lib/apis/audio';
-	import { user, settings, config } from '$lib/stores';
-	import { createEventDispatcher, onMount, getContext } from 'svelte';
 	import { toast } from 'svelte-sonner';
-	import Switch from '$lib/components/common/Switch.svelte';
+	import { createEventDispatcher, onMount, getContext } from 'svelte';
+	const dispatch = createEventDispatcher();
+
 	import { getBackendConfig } from '$lib/apis';
+	import {
+		getAudioConfig,
+		updateAudioConfig,
+		getModels as _getModels,
+		getVoices as _getVoices
+	} from '$lib/apis/audio';
+	import { user, settings, config } from '$lib/stores';
+
 	import SensitiveInput from '$lib/components/common/SensitiveInput.svelte';
-	const dispatch = createEventDispatcher();
 
 	const i18n = getContext('i18n');
 
@@ -30,22 +36,23 @@
 	let models = [];
 	let nonLocalVoices = false;
 
-	const getOpenAIVoices = () => {
-		voices = [
-			{ name: 'alloy' },
-			{ name: 'echo' },
-			{ name: 'fable' },
-			{ name: 'onyx' },
-			{ name: 'nova' },
-			{ name: 'shimmer' }
-		];
-	};
+	const getModels = async () => {
+		if (TTS_ENGINE === '') {
+			models = [];
+		} else {
+			const res = await _getModels(localStorage.token).catch((e) => {
+				toast.error(e);
+			});
 
-	const getOpenAIModels = () => {
-		models = [{ name: 'tts-1' }, { name: 'tts-1-hd' }];
+			if (res) {
+				console.log(res);
+				models = res.models;
+			}
+		}
 	};
 
-	const getWebAPIVoices = () => {
+	const getVoices = async () => {
+		if (TTS_ENGINE === '') {
 			const getVoicesLoop = setInterval(async () => {
 				voices = await speechSynthesis.getVoices();
 
@@ -54,6 +61,16 @@
 					clearInterval(getVoicesLoop);
 				}
 			}, 100);
+		} else {
+			const res = await _getVoices(localStorage.token).catch((e) => {
+				toast.error(e);
+			});
+
+			if (res) {
+				console.log(res);
+				voices = res.voices;
+			}
+		}
 	};
 
 	const updateConfigHandler = async () => {
@@ -101,12 +118,8 @@
 			STT_MODEL = res.stt.MODEL;
 		}
 
-		if (TTS_ENGINE === 'openai') {
-			getOpenAIVoices();
-			getOpenAIModels();
-		} else {
-			getWebAPIVoices();
-		}
+		await getVoices();
+		await getModels();
 	});
 </script>
 
@@ -185,13 +198,15 @@
 							class=" dark:bg-gray-900 w-fit pr-8 rounded px-2 p-1 text-xs bg-transparent outline-none text-right"
 							bind:value={TTS_ENGINE}
 							placeholder="Select a mode"
-							on:change={(e) => {
+							on:change={async (e) => {
+								await updateConfigHandler();
+								await getVoices();
+								await getModels();
+
 								if (e.target.value === 'openai') {
-									getOpenAIVoices();
 									TTS_VOICE = 'alloy';
 									TTS_MODEL = 'tts-1';
 								} else {
-									getWebAPIVoices();
 									TTS_VOICE = '';
 									TTS_MODEL = '';
 								}
@@ -268,7 +283,7 @@
 
 									<datalist id="voice-list">
 										{#each voices as voice}
-											<option value={voice.name} />
+											<option value={voice.id}>{voice.name}</option>
 										{/each}
 									</datalist>
 								</div>
@@ -279,15 +294,15 @@
 							<div class="flex w-full">
 								<div class="flex-1">
 									<input
-										list="model-list"
+										list="tts-model-list"
 										class="w-full rounded-lg py-2 px-4 text-sm bg-gray-50 dark:text-gray-300 dark:bg-gray-850 outline-none"
 										bind:value={TTS_MODEL}
 										placeholder="Select a model"
 									/>
 
-									<datalist id="model-list">
+									<datalist id="tts-model-list">
 										{#each models as model}
-											<option value={model.name} />
+											<option value={model.id} />
 										{/each}
 									</datalist>
 								</div>
@@ -309,7 +324,7 @@
 
 									<datalist id="voice-list">
 										{#each voices as voice}
-											<option value={voice.name} />
+											<option value={voice.id}>{voice.name}</option>
 										{/each}
 									</datalist>
 								</div>
@@ -320,15 +335,15 @@
 							<div class="flex w-full">
 								<div class="flex-1">
 									<input
-										list="model-list"
+										list="tts-model-list"
 										class="w-full rounded-lg py-2 px-4 text-sm bg-gray-50 dark:text-gray-300 dark:bg-gray-850 outline-none"
 										bind:value={TTS_MODEL}
 										placeholder="Select a model"
 									/>
 
-									<datalist id="model-list">
+									<datalist id="tts-model-list">
 										{#each models as model}
-											<option value={model.name} />
+											<option value={model.id} />
 										{/each}
 									</datalist>
 								</div>

src/lib/components/chat/Settings/Audio.svelte

 <script lang="ts">
-	import { user, settings, config } from '$lib/stores';
-	import { createEventDispatcher, onMount, getContext } from 'svelte';
 	import { toast } from 'svelte-sonner';
+	import { createEventDispatcher, onMount, getContext } from 'svelte';
+
+	import { user, settings, config } from '$lib/stores';
+	import { getVoices as _getVoices } from '$lib/apis/audio';
+
 	import Switch from '$lib/components/common/Switch.svelte';
 	const dispatch = createEventDispatcher();
 
@@ -20,18 +23,8 @@
 	let voices = [];
 	let voice = '';
 
-	const getOpenAIVoices = () => {
-		voices = [
-			{ name: 'alloy' },
-			{ name: 'echo' },
-			{ name: 'fable' },
-			{ name: 'onyx' },
-			{ name: 'nova' },
-			{ name: 'shimmer' }
-		];
-	};
-
-	const getWebAPIVoices = () => {
+	const getVoices = async () => {
+		if ($config.audio.tts.engine === '') {
 			const getVoicesLoop = setInterval(async () => {
 				voices = await speechSynthesis.getVoices();
 
@@ -40,6 +33,16 @@
 					clearInterval(getVoicesLoop);
 				}
 			}, 100);
+		} else {
+			const res = await _getVoices(localStorage.token).catch((e) => {
+				toast.error(e);
+			});
+
+			if (res) {
+				console.log(res);
+				voices = res.voices;
+			}
+		}
 	};
 
 	const toggleResponseAutoPlayback = async () => {
@@ -61,11 +64,7 @@
 		voice = $settings?.audio?.tts?.voice ?? $config.audio.tts.voice ?? '';
 		nonLocalVoices = $settings.audio?.tts?.nonLocalVoices ?? false;
 
-		if ($config.audio.tts.engine === 'openai') {
-			getOpenAIVoices();
-		} else {
-			getWebAPIVoices();
-		}
+		await getVoices();
 	});
 </script>
 
@@ -195,7 +194,7 @@
 
 						<datalist id="voice-list">
 							{#each voices as voice}
-								<option value={voice.name} />
+								<option value={voice.id}>{voice.name}</option>
 							{/each}
 						</datalist>
 					</div>