Narration
Consistent long-form voice.
Use a single cloned narrator across an entire audiobook, course, or documentary, no drift between chapters.
Voice cloning
Clone a narrator, an executive, or a brand persona from a 10-second sample. Instant; consent-gated; dialect-aware; never trained on anyone who didn't opt in.
How it works
Consent, mint, deployment, each step gated by policy, not by engineering delay.
Upload 10 seconds of clean audio with a signed consent attestation for the speaker.
Cloning is instant. The voice id is ready in the same response, no training queue, no email to wait for.
Swap the `voice` id into any `Hakim Fast v1.3` call. TTS, streaming, batch, all immediate.
Try it
Cloning is consent-gated, so it's never anonymous-safe. Create a free account and step through the wizard.
Voice cloning requires a consent attestation for every sample. Sign up to step through the guided wizard in the playground.
Where teams ship this
Every clone on Hakim is consent-attested and scoped to the workspace that owns the samples.
Narration
Use a single cloned narrator across an entire audiobook, course, or documentary, no drift between chapters.
Brand voice
IVR, ads, explainer videos, product tours, one custom voice, delivered through one API.
Accessibility
Clinicians and families can capture a speaker's voice while they're able, and keep using it as a communication aid.
Technical specs
API
Once a voice is cloned, its id plugs straight into every Hakim Fast v1.3 call you already make.
import { Hakim } from '@hakim/sdk-node';
const client = new Hakim({ apiKey: process.env.HAKIM_API_KEY });
const voice = await client.voices.clones.create({
name: 'Brand voice (Khaleeji)',
language: 'ar',
dialect: 'khaleeji',
consent_attested: true,
samples: ['sample-1.wav', 'sample-2.wav', 'sample-3.wav'],
});
console.log(voice.id);See the full reference in the docs.
Ready to clone?
Free accounts can clone one voice to try the flow. Move to Creator for commercial use.