MIGRATING.md

Migrating from v1 to v2

Client Initialization

// In both v1 and v2, the API key can be retrieved using the LMNT_API_KEY environment variable or passed as an argument as below.
// v1
import Lmnt from 'lmnt-node';
const client = new Lmnt('your-api-key');

// v2
import Lmnt from 'lmnt-node';
const client = new Lmnt({ apiKey: 'your-api-key' });

Synthesis

// v1
const result = await client.synthesize('Hello world', 'voice-id'); // result: { audio: Buffer }

// v2
const response = await client.speech.generate({
    text: 'Hello world',
    voice: 'voice-id',
}); // result: Response
// To get the data all at once:
const buffer = Buffer.from(await response.arrayBuffer());
// Or to stream the data:
for await (const chunk of response.body) {
    // chunk is a Uint8Array of audio data
}

Streaming (Speech Session)

// v1
const stream = client.synthesizeStreaming('voice-id', {
    return_extras: true
});
stream.appendText('Hello world');
stream.finish();
for await (const chunk of stream) {
    // chunk: { audio: Buffer, durations?: Duration[], warning?: string }
}

// v2
const session = await client.speech.sessions.create({
    voice: 'voice-id',
    return_extras: true
});
session.appendText('Hello world');
session.finish();
for await (const chunk of session) {
    // chunk: { audio: Buffer, durations?: Duration[], warning?: string }
}

Voice Management

// v1
const voices = await client.fetchVoices({ starred: true });
const voice = await client.fetchVoice('voice-id');
await client.updateVoice('voice-id', { name: 'New Name' });
await client.deleteVoice('voice-id');
const filenames = ['filename1', 'filename2'];
await client.createVoice('voice-name', False, filenames)
// v2
const voices = await client.voices.list({ starred: true });
const voice = await client.voices.retrieve('voice-id');
await client.voices.update('voice-id', { name: 'New Name' });
await client.voices.delete('voice-id');
const files = [fs.createReadStream('filename1'), fs.createReadStream('filename2')];
await client.voices.create('voice-name', False, files)

Account Information

// v1
const account = await client.fetchAccount();
// v2
const account = await client.accounts.retrieve();

Error Handling

The SpeechError class has been replaced with standard error types from the SDK. Errors will contain appropriate status codes and error messages.

Runtime Detection

The v1 SDK's runtime detection for Node.js vs browser environments has been replaced with built-in environment handling in v2. You no longer need to handle this manually.