@micdrop/elevenlabs 1.2.1 → 1.4.0
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- package/README.md +9 -8
- package/dist/index.d.mts +46 -3
- package/dist/index.d.ts +46 -3
- package/dist/index.js +419 -388
- package/dist/index.js.map +1 -1
- package/dist/index.mjs +421 -377
- package/dist/index.mjs.map +1 -1
- package/package.json +2 -2
package/README.md
CHANGED
|
@@ -1,6 +1,6 @@
|
|
|
1
1
|
# @micdrop/elevenlabs
|
|
2
2
|
|
|
3
|
-
[Micdrop website](https://micdrop.dev) | [Documentation](https://micdrop.dev/docs/
|
|
3
|
+
[Micdrop website](https://micdrop.dev) | [Documentation](https://micdrop.dev/docs/text-to-speech/elevenlabs)
|
|
4
4
|
|
|
5
5
|
`@micdrop/elevenlabs` streams ElevenLabs text to speech for [@micdrop/server](https://micdrop.dev/docs/server). It works with Eleven v4 Turbo, Eleven v4, Eleven v3, Flash v2.5, Turbo v2.5 and Multilingual v2, and picks the WebSocket each model needs.
|
|
6
6
|
|
|
@@ -23,6 +23,7 @@ const tts = new ElevenLabsTTS({
|
|
|
23
23
|
voiceId: '21m00Tcm4TlvDq8ikWAM', // ElevenLabs voice ID
|
|
24
24
|
modelId: 'eleven_v4_turbo', // Optional: model to use
|
|
25
25
|
language: 'en', // Optional: language code
|
|
26
|
+
sampleRate: 24000, // Optional: 48000 by default
|
|
26
27
|
voiceSettings: {
|
|
27
28
|
stability: 0.5,
|
|
28
29
|
},
|
|
@@ -46,7 +47,7 @@ const tts = new ElevenLabsTTS({
|
|
|
46
47
|
voiceId: '21m00Tcm4TlvDq8ikWAM',
|
|
47
48
|
})
|
|
48
49
|
|
|
49
|
-
// Audio is raw PCM, 16 bits,
|
|
50
|
+
// Audio is raw PCM, 16 bits, mono, at tts.sampleRate
|
|
50
51
|
tts.on('Audio', (chunk) => console.log('Audio:', chunk.length, 'bytes'))
|
|
51
52
|
tts.on('Failed', (texts) => console.error('Failed:', texts))
|
|
52
53
|
|
|
@@ -55,12 +56,12 @@ tts.speak(Readable.from(['Hello! ', 'What can I do for you?']))
|
|
|
55
56
|
|
|
56
57
|
### Events
|
|
57
58
|
|
|
58
|
-
| Event | Payload | Description
|
|
59
|
-
| -------- | ---------- |
|
|
60
|
-
| `Audio` | `Buffer` | A chunk of audio, PCM 16 bits,
|
|
61
|
-
| `Failed` | `string[]` | The text left unspoken when `ElevenLabsTTS` gives up after its retries.
|
|
59
|
+
| Event | Payload | Description |
|
|
60
|
+
| -------- | ---------- | -------------------------------------------------------------------------------------------------------------------------- |
|
|
61
|
+
| `Audio` | `Buffer` | A chunk of audio, PCM 16 bits, mono, at `tts.sampleRate` (the `sampleRate` option, 48 kHz by default), ready to be played. |
|
|
62
|
+
| `Failed` | `string[]` | The text left unspoken when `ElevenLabsTTS` gives up after its retries. |
|
|
62
63
|
|
|
63
|
-
See the [TTS interface](https://micdrop.dev/docs/
|
|
64
|
+
See the [TTS interface](https://micdrop.dev/docs/text-to-speech/custom) for the full contract.
|
|
64
65
|
|
|
65
66
|
### Models
|
|
66
67
|
|
|
@@ -77,7 +78,7 @@ ElevenLabs counts each open connection to the Text to Dialogue WebSocket as one
|
|
|
77
78
|
|
|
78
79
|
## Documentation
|
|
79
80
|
|
|
80
|
-
Read full [documentation of the ElevenLabs integration for Micdrop](https://micdrop.dev/docs/
|
|
81
|
+
Read full [documentation of the ElevenLabs integration for Micdrop](https://micdrop.dev/docs/text-to-speech/elevenlabs) on the [website](https://micdrop.dev).
|
|
81
82
|
|
|
82
83
|
## License
|
|
83
84
|
|
package/dist/index.d.mts
CHANGED
|
@@ -1,13 +1,24 @@
|
|
|
1
|
-
import { TTS } from '@micdrop/server';
|
|
1
|
+
import { TTS, MicdropModel, MicdropVoice } from '@micdrop/server';
|
|
2
2
|
import { Readable } from 'stream';
|
|
3
3
|
import { TextToSpeechStreamRequestOutputFormat, VoiceSettings } from '@elevenlabs/elevenlabs-js/api';
|
|
4
4
|
|
|
5
5
|
type ElevenLabsModelId = 'eleven_v4_turbo' | 'eleven_v4' | 'eleven_v3_conversational' | 'eleven_v3' | 'eleven_flash_v2_5' | 'eleven_multilingual_v2' | 'eleven_turbo_v2_5';
|
|
6
|
+
/** Rates ElevenLabs streams raw PCM at. 44100 needs a Pro plan or above. */
|
|
7
|
+
type ElevenLabsSampleRate = 8000 | 16000 | 22050 | 24000 | 32000 | 44100 | 48000;
|
|
6
8
|
interface ElevenLabsTTSOptions {
|
|
7
9
|
apiKey: string;
|
|
8
10
|
voiceId: string;
|
|
9
11
|
modelId?: ElevenLabsModelId;
|
|
10
12
|
language?: string;
|
|
13
|
+
/**
|
|
14
|
+
* Rate of the audio, in Hz, 48000 by default, the highest one offered. It
|
|
15
|
+
* goes to the client as is, so the voice is never resampled on the server.
|
|
16
|
+
*/
|
|
17
|
+
sampleRate?: ElevenLabsSampleRate;
|
|
18
|
+
/**
|
|
19
|
+
* @deprecated Use `sampleRate`. The client only plays raw PCM, so only the
|
|
20
|
+
* `pcm_*` formats work.
|
|
21
|
+
*/
|
|
11
22
|
outputFormat?: TextToSpeechStreamRequestOutputFormat;
|
|
12
23
|
voiceSettings?: VoiceSettings;
|
|
13
24
|
connectionTimeout?: number;
|
|
@@ -15,7 +26,13 @@ interface ElevenLabsTTSOptions {
|
|
|
15
26
|
maxRetry?: number;
|
|
16
27
|
}
|
|
17
28
|
declare const DEFAULT_MODEL_ID = "eleven_turbo_v2_5";
|
|
18
|
-
declare const
|
|
29
|
+
declare const DEFAULT_SAMPLE_RATE: ElevenLabsSampleRate;
|
|
30
|
+
/** @deprecated Use `DEFAULT_SAMPLE_RATE` */
|
|
31
|
+
declare const DEFAULT_OUTPUT_FORMAT = "pcm_48000";
|
|
32
|
+
/** The `output_format` asked of ElevenLabs */
|
|
33
|
+
declare function getOutputFormat(options: ElevenLabsTTSOptions): TextToSpeechStreamRequestOutputFormat;
|
|
34
|
+
/** Rate of the audio ElevenLabs sends, read from its `output_format` */
|
|
35
|
+
declare function getSampleRate(options: ElevenLabsTTSOptions): number;
|
|
19
36
|
/** Eleven v3 and v4 only run on the Text to Dialogue WebSocket */
|
|
20
37
|
declare function isDialogueModel(modelId: string): boolean;
|
|
21
38
|
type ElevenLabsWebSocketMessage = ElevenLabsWebSocketAudioOutputMessage | ElevenLabsWebSocketFinalOutputMessage | ElevenLabsWebSocketErrorMessage;
|
|
@@ -68,11 +85,37 @@ type ElevenLabsDialogueMessage = {
|
|
|
68
85
|
* so the model picks the one this class drives.
|
|
69
86
|
*/
|
|
70
87
|
declare class ElevenLabsTTS extends TTS {
|
|
88
|
+
/**
|
|
89
|
+
* Lists the models that speak, with the languages each one reads.
|
|
90
|
+
*
|
|
91
|
+
* The list is fetched once and shared by every call of the process for a
|
|
92
|
+
* day. `refresh` fetches it again.
|
|
93
|
+
*/
|
|
94
|
+
static listModels(options: {
|
|
95
|
+
apiKey: string;
|
|
96
|
+
refresh?: boolean;
|
|
97
|
+
}): Promise<MicdropModel[]>;
|
|
98
|
+
/**
|
|
99
|
+
* Lists the voices the account can speak with: the premade ones, and the
|
|
100
|
+
* ones cloned, designed or added from the library.
|
|
101
|
+
*
|
|
102
|
+
* `language`, such as `fr` or `fr-FR`, keeps the voices that speak it.
|
|
103
|
+
*
|
|
104
|
+
* The list is fetched once and shared by every call of the process for a
|
|
105
|
+
* day. `refresh` fetches it again, for a voice added a moment ago.
|
|
106
|
+
*/
|
|
107
|
+
static listVoices(options: {
|
|
108
|
+
apiKey: string;
|
|
109
|
+
language?: string;
|
|
110
|
+
refresh?: boolean;
|
|
111
|
+
}): Promise<MicdropVoice[]>;
|
|
71
112
|
private readonly webSocket;
|
|
113
|
+
private readonly rate;
|
|
72
114
|
constructor(options: ElevenLabsTTSOptions);
|
|
115
|
+
get sampleRate(): number;
|
|
73
116
|
speak(textStream: Readable): void;
|
|
74
117
|
cancel(): void;
|
|
75
118
|
destroy(): void;
|
|
76
119
|
}
|
|
77
120
|
|
|
78
|
-
export { DEFAULT_MODEL_ID, DEFAULT_OUTPUT_FORMAT, type ElevenLabsDialogueClientMessage, type ElevenLabsDialogueInput, type ElevenLabsDialogueMessage, type ElevenLabsModelId, ElevenLabsTTS, type ElevenLabsTTSOptions, type ElevenLabsWebSocketAudioOutputMessage, type ElevenLabsWebSocketAudioOutputMessageAlignment, type ElevenLabsWebSocketErrorMessage, type ElevenLabsWebSocketFinalOutputMessage, type ElevenLabsWebSocketMessage, isDialogueModel };
|
|
121
|
+
export { DEFAULT_MODEL_ID, DEFAULT_OUTPUT_FORMAT, DEFAULT_SAMPLE_RATE, type ElevenLabsDialogueClientMessage, type ElevenLabsDialogueInput, type ElevenLabsDialogueMessage, type ElevenLabsModelId, type ElevenLabsSampleRate, ElevenLabsTTS, type ElevenLabsTTSOptions, type ElevenLabsWebSocketAudioOutputMessage, type ElevenLabsWebSocketAudioOutputMessageAlignment, type ElevenLabsWebSocketErrorMessage, type ElevenLabsWebSocketFinalOutputMessage, type ElevenLabsWebSocketMessage, getOutputFormat, getSampleRate, isDialogueModel };
|
package/dist/index.d.ts
CHANGED
|
@@ -1,13 +1,24 @@
|
|
|
1
|
-
import { TTS } from '@micdrop/server';
|
|
1
|
+
import { TTS, MicdropModel, MicdropVoice } from '@micdrop/server';
|
|
2
2
|
import { Readable } from 'stream';
|
|
3
3
|
import { TextToSpeechStreamRequestOutputFormat, VoiceSettings } from '@elevenlabs/elevenlabs-js/api';
|
|
4
4
|
|
|
5
5
|
type ElevenLabsModelId = 'eleven_v4_turbo' | 'eleven_v4' | 'eleven_v3_conversational' | 'eleven_v3' | 'eleven_flash_v2_5' | 'eleven_multilingual_v2' | 'eleven_turbo_v2_5';
|
|
6
|
+
/** Rates ElevenLabs streams raw PCM at. 44100 needs a Pro plan or above. */
|
|
7
|
+
type ElevenLabsSampleRate = 8000 | 16000 | 22050 | 24000 | 32000 | 44100 | 48000;
|
|
6
8
|
interface ElevenLabsTTSOptions {
|
|
7
9
|
apiKey: string;
|
|
8
10
|
voiceId: string;
|
|
9
11
|
modelId?: ElevenLabsModelId;
|
|
10
12
|
language?: string;
|
|
13
|
+
/**
|
|
14
|
+
* Rate of the audio, in Hz, 48000 by default, the highest one offered. It
|
|
15
|
+
* goes to the client as is, so the voice is never resampled on the server.
|
|
16
|
+
*/
|
|
17
|
+
sampleRate?: ElevenLabsSampleRate;
|
|
18
|
+
/**
|
|
19
|
+
* @deprecated Use `sampleRate`. The client only plays raw PCM, so only the
|
|
20
|
+
* `pcm_*` formats work.
|
|
21
|
+
*/
|
|
11
22
|
outputFormat?: TextToSpeechStreamRequestOutputFormat;
|
|
12
23
|
voiceSettings?: VoiceSettings;
|
|
13
24
|
connectionTimeout?: number;
|
|
@@ -15,7 +26,13 @@ interface ElevenLabsTTSOptions {
|
|
|
15
26
|
maxRetry?: number;
|
|
16
27
|
}
|
|
17
28
|
declare const DEFAULT_MODEL_ID = "eleven_turbo_v2_5";
|
|
18
|
-
declare const
|
|
29
|
+
declare const DEFAULT_SAMPLE_RATE: ElevenLabsSampleRate;
|
|
30
|
+
/** @deprecated Use `DEFAULT_SAMPLE_RATE` */
|
|
31
|
+
declare const DEFAULT_OUTPUT_FORMAT = "pcm_48000";
|
|
32
|
+
/** The `output_format` asked of ElevenLabs */
|
|
33
|
+
declare function getOutputFormat(options: ElevenLabsTTSOptions): TextToSpeechStreamRequestOutputFormat;
|
|
34
|
+
/** Rate of the audio ElevenLabs sends, read from its `output_format` */
|
|
35
|
+
declare function getSampleRate(options: ElevenLabsTTSOptions): number;
|
|
19
36
|
/** Eleven v3 and v4 only run on the Text to Dialogue WebSocket */
|
|
20
37
|
declare function isDialogueModel(modelId: string): boolean;
|
|
21
38
|
type ElevenLabsWebSocketMessage = ElevenLabsWebSocketAudioOutputMessage | ElevenLabsWebSocketFinalOutputMessage | ElevenLabsWebSocketErrorMessage;
|
|
@@ -68,11 +85,37 @@ type ElevenLabsDialogueMessage = {
|
|
|
68
85
|
* so the model picks the one this class drives.
|
|
69
86
|
*/
|
|
70
87
|
declare class ElevenLabsTTS extends TTS {
|
|
88
|
+
/**
|
|
89
|
+
* Lists the models that speak, with the languages each one reads.
|
|
90
|
+
*
|
|
91
|
+
* The list is fetched once and shared by every call of the process for a
|
|
92
|
+
* day. `refresh` fetches it again.
|
|
93
|
+
*/
|
|
94
|
+
static listModels(options: {
|
|
95
|
+
apiKey: string;
|
|
96
|
+
refresh?: boolean;
|
|
97
|
+
}): Promise<MicdropModel[]>;
|
|
98
|
+
/**
|
|
99
|
+
* Lists the voices the account can speak with: the premade ones, and the
|
|
100
|
+
* ones cloned, designed or added from the library.
|
|
101
|
+
*
|
|
102
|
+
* `language`, such as `fr` or `fr-FR`, keeps the voices that speak it.
|
|
103
|
+
*
|
|
104
|
+
* The list is fetched once and shared by every call of the process for a
|
|
105
|
+
* day. `refresh` fetches it again, for a voice added a moment ago.
|
|
106
|
+
*/
|
|
107
|
+
static listVoices(options: {
|
|
108
|
+
apiKey: string;
|
|
109
|
+
language?: string;
|
|
110
|
+
refresh?: boolean;
|
|
111
|
+
}): Promise<MicdropVoice[]>;
|
|
71
112
|
private readonly webSocket;
|
|
113
|
+
private readonly rate;
|
|
72
114
|
constructor(options: ElevenLabsTTSOptions);
|
|
115
|
+
get sampleRate(): number;
|
|
73
116
|
speak(textStream: Readable): void;
|
|
74
117
|
cancel(): void;
|
|
75
118
|
destroy(): void;
|
|
76
119
|
}
|
|
77
120
|
|
|
78
|
-
export { DEFAULT_MODEL_ID, DEFAULT_OUTPUT_FORMAT, type ElevenLabsDialogueClientMessage, type ElevenLabsDialogueInput, type ElevenLabsDialogueMessage, type ElevenLabsModelId, ElevenLabsTTS, type ElevenLabsTTSOptions, type ElevenLabsWebSocketAudioOutputMessage, type ElevenLabsWebSocketAudioOutputMessageAlignment, type ElevenLabsWebSocketErrorMessage, type ElevenLabsWebSocketFinalOutputMessage, type ElevenLabsWebSocketMessage, isDialogueModel };
|
|
121
|
+
export { DEFAULT_MODEL_ID, DEFAULT_OUTPUT_FORMAT, DEFAULT_SAMPLE_RATE, type ElevenLabsDialogueClientMessage, type ElevenLabsDialogueInput, type ElevenLabsDialogueMessage, type ElevenLabsModelId, type ElevenLabsSampleRate, ElevenLabsTTS, type ElevenLabsTTSOptions, type ElevenLabsWebSocketAudioOutputMessage, type ElevenLabsWebSocketAudioOutputMessageAlignment, type ElevenLabsWebSocketErrorMessage, type ElevenLabsWebSocketFinalOutputMessage, type ElevenLabsWebSocketMessage, getOutputFormat, getSampleRate, isDialogueModel };
|