@absolutejs/voice-smallest

Smallest AI text-to-speech adapter for @absolutejs/voice.

Wraps Smallest AI's /api/v1/{model}/get_speech HTTP endpoint behind voice's TTSAdapter seam. Lightning + Lightning-v2 models, English + Hindi.

Install

bun add @absolutejs/voice-smallest

@absolutejs/voice is a runtime dependency.

Use

import { voice } from "@absolutejs/voice";
import { smallest } from "@absolutejs/voice-smallest";

const app = voice({
  tts: smallest({
    apiKey: process.env.SMALLEST_API_KEY!,
    voiceId: "george",
    // optional:
    model: "lightning",          // default; or 'lightning-v2'
    sampleRate: 24_000,          // default
    language: "en",
    speed: 1.0,
    similarity: 0.5,
    consistency: 0.5,
    enhancement: 1,
  }),
});

Options

| Option | Required | Default | Notes | | --- | --- | --- | --- | | apiKey | yes | — | Smallest API key, sent as Authorization: Bearer <key>. | | voiceId | yes | — | Smallest voice id. | | model | no | lightning | Selects the URL path segment (/api/v1/{model}/get_speech). | | sampleRate | no | 24_000 | Forwarded as sample_rate. Adapter emits PCM s16le frames at this rate. | | language | no | — | en, hi. | | speed, similarity, consistency, enhancement | no | — | Forwarded to Smallest. | | baseUrl | no | https://waves-api.smallest.ai | Override for staging. | | fetch | no | globalThis.fetch | Inject for tests; opportunistic HTTP/2 multiplexing on outbound HTTPS. |

Notes

• Whitespace-only send() is a no-op.
• Adapter requests add_wav_header: false so the body is raw PCM.
• session.close(reason) aborts in-flight requests and refuses further sends.