Persona Voice Cloning → Branded Audio

Scenario

Clone a spokesperson’s voice using ElevenLabs’ voice cloning API, then generate audio versions of Mavera-generated content using that cloned voice — building a consistent audio identity across all content.

Flow: ElevenLabs POST /voices/add → get voice_id → Mavera POST /generations → ElevenLabs POST /text-to-speech/{voice_id} → branded audio

Code

Python

import os, requests, time
EL_KEY = os.environ["ELEVENLABS_API_KEY"]
EL_BASE = "https://api.elevenlabs.io/v1"
MV = os.environ["MAVERA_API_KEY"]
MV_BASE = "https://app.mavera.io/api/v1"
MV_H = {"Authorization": f"Bearer {MV}", "Content-Type": "application/json"}
os.makedirs("branded_audio", exist_ok=True)
SAMPLE_FILES = ["voice_samples/intro.mp3", "voice_samples/pitch.mp3", "voice_samples/casual.mp3"]

# 1. Clone voice with multipart upload
files = [("files", (os.path.basename(f), open(f, "rb"), "audio/mpeg")) for f in SAMPLE_FILES]
clone = requests.post(f"{EL_BASE}/voices/add", headers={"xi-api-key": EL_KEY},
    data={"name": "Brand Spokesperson v1",
          "description": "Official brand voice from spokesperson recordings",
          "labels": '{"use_case": "marketing", "language": "en"}'},
    files=files).json()
for _, (_, fh, _) in zip(SAMPLE_FILES, files): fh.close()
voice_id = clone["voice_id"]
print(f"Voice cloned: {voice_id}")
time.sleep(2)

# 2. Generate content + convert to branded audio
SPECS = [
    {"prompt": "Write a 100-word product announcement for an AI analytics dashboard.", "label": "product-launch"},
    {"prompt": "Write a 60-second radio ad script for a B2B marketing platform.", "label": "radio-spot"},
    {"prompt": "Write a 30-second podcast pre-roll ad for a data intelligence tool.", "label": "podcast-preroll"},
]
for spec in SPECS:
    gen = requests.post(f"{MV_BASE}/generations", headers=MV_H,
        json={"prompt": spec["prompt"]}).json()
    text = gen.get("output") or gen.get("content") or ""
    print(f"[{spec['label']}] {len(text.split())} words")
    time.sleep(1)

    tts = requests.post(f"{EL_BASE}/text-to-speech/{voice_id}",
        headers={"xi-api-key": EL_KEY, "Content-Type": "application/json"},
        json={"text": text, "model_id": "eleven_multilingual_v2",
              "voice_settings": {"stability": 0.5, "similarity_boost": 0.8,
                                 "style": 0.3, "use_speaker_boost": True}})
    if tts.status_code == 200:
        path = f"branded_audio/{spec['label']}-cloned.mp3"
        with open(path, "wb") as f: f.write(tts.content)
        print(f"  → {path} ({len(tts.content) // 1024} KB)")
    time.sleep(2)

import fs from "fs";
import path from "path";

const EL_KEY = process.env.ELEVENLABS_API_KEY;
const EL_BASE = "https://api.elevenlabs.io/v1";
const MV = process.env.MAVERA_API_KEY;
const MV_BASE = "https://app.mavera.io/api/v1";
const MV_H = { Authorization: `Bearer ${MV}`, "Content-Type": "application/json" };
fs.mkdirSync("branded_audio", { recursive: true });

const sampleFiles = ["voice_samples/intro.mp3", "voice_samples/pitch.mp3", "voice_samples/casual.mp3"];

// 1. Clone voice
const form = new FormData();
form.append("name", "Brand Spokesperson v1");
form.append("description", "Official brand voice from spokesperson recordings");
form.append("labels", JSON.stringify({ use_case: "marketing", language: "en" }));
for (const fp of sampleFiles)
  form.append("files", new Blob([fs.readFileSync(fp)]), path.basename(fp));

const clone = await fetch(`${EL_BASE}/voices/add`, {
  method: "POST", headers: { "xi-api-key": EL_KEY }, body: form,
}).then(r => r.json());
const voiceId = clone.voice_id;
console.log(`Voice cloned: ${voiceId}`);
await new Promise(r => setTimeout(r, 2000));

// 2. Generate content + audio
const specs = [
  { prompt: "Write a 100-word product announcement for an AI analytics dashboard.", label: "product-launch" },
  { prompt: "Write a 60-second radio ad for a B2B marketing platform.", label: "radio-spot" },
  { prompt: "Write a 30-second podcast pre-roll ad.", label: "podcast-preroll" },
];
for (const spec of specs) {
  const gen = await fetch(`${MV_BASE}/generations`, { method: "POST", headers: MV_H,
    body: JSON.stringify({ prompt: spec.prompt }) }).then(r => r.json());
  const text = gen.output || gen.content || "";
  console.log(`[${spec.label}] ${text.split(/\s+/).length} words`);
  await new Promise(r => setTimeout(r, 1000));

  const tts = await fetch(`${EL_BASE}/text-to-speech/${voiceId}`, { method: "POST",
    headers: { "xi-api-key": EL_KEY, "Content-Type": "application/json" },
    body: JSON.stringify({ text, model_id: "eleven_multilingual_v2",
      voice_settings: { stability: 0.5, similarity_boost: 0.8, style: 0.3, use_speaker_boost: true } }),
  });
  if (tts.ok) {
    const buf = Buffer.from(await tts.arrayBuffer());
    const out = `branded_audio/${spec.label}-cloned.mp3`;
    fs.writeFileSync(out, buf);
    console.log(`  → ${out} (${Math.round(buf.length / 1024)} KB)`);
  }
  await new Promise(r => setTimeout(r, 2000));
}

Example Output

Voice cloned: vId_a8f3c2d1e9b7
[product-launch] 104 words → product-launch-cloned.mp3 (142 KB)
[radio-spot] 87 words → radio-spot-cloned.mp3 (198 KB)
[podcast-preroll] 62 words → podcast-preroll-cloned.mp3 (89 KB)

Error Handling

Voice cloning sample requirements

Instant cloning accepts 1–25 samples. Provide at least 3 clean samples totaling 1–3 minutes with only the target speaker — no background music or multiple speakers.

Character credit limits

Each plan has a monthly character quota (~500 chars per 100 words). A 401 with quota_exceeded means your allocation is exhausted — upgrade or wait for monthly reset.

Voice not found after cloning

If TTS returns 404 after cloning, the voice may still be processing. Wait 5–10 seconds and retry. Verify with GET /voices.