HeyGen Voice Turbo: is it free, and how to use it

Yes, HeyGen Voice Turbo is free right now: the heygen-voice-1-turbo-free model costs nothing through HeyGen's API until November 30, 2026, and you don't need a credit card. After that, HeyGen's own pricing page lists the paid Turbo model at $4 per million characters, with public voices and instant clones of your own voice.

The posts that put it on the map

On October 21, @HeyGen announced that it held the top two spots on the Artificial Analysis text-to-speech leaderboard: Voice Turbo first, the standard HeyGen Voice second. The company calls Turbo the faster, cheaper mode of the voice model behind every HeyGen video, and says the endpoint is free until the end of November.

Less than a minute later, @ArtificialAnlys posted the numbers. Turbo "takes the #1 spot" in its Controlled Voice arena with an Elo of 1,202 across 1,663 blind matchups, first of 46 models. It also places first for UK English, second for US English behind Qwen-Audio-3.1-TTS-Plus, and first at reading exact sequences such as numbers and codes correctly (91.8%). On general pronunciation it is eighth of 33, so it doesn't lead every category.

Overnight, @AGTPinsights put it next to the competition, saying HeyGen's original announcement had passed 819K views. Its list includes Eleven v4 Turbo, Voxtral 3 TTS and a Gemini Flash-Lite voice model, all more expensive per character.

What it costs

The prices don't match across sources, so this table uses HeyGen's own launch page:

Artificial Analysis lists the standard model at $30, and AGTP's post puts Turbo at $7.50. One of those is likely an estimate or a reseller rate. Either way, check the billing page of whichever provider you sign up with before you plan a budget around a number.

The free tier runs on shared capacity. HeyGen's docs say there's no guarantee on uptime or time to first audio, and a busy free queue returns HTTP 429. A free request never quietly switches to a paid model, which is nice to know before you leave a script running overnight.

How to try it

1. Make a HeyGen account and create an API key at platform.heygen.com/api-keys.
2. Listen to voices in the playground at platform.heygen.com/models/heygen-voice, or list the public ones with GET /v3/models/audio/voices?type=public. Each voice includes a preview clip.
3. Send text to the speech endpoint. Here's HeyGen's sample request:

curl -X POST "https://api.heygen.com/v3/models/audio/tts" \
  -H "X-Api-Key: $HEYGEN_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "heygen-voice-1-turbo-free",
    "voice_id": "83548de556df48ba8c09c42a57c51d85",
    "text": "Hello from HeyGen Voice Turbo."
  }'

The response links to a WAV file. There's also a /tts/stream endpoint that sends audio in pieces as it's generated, which is the one to use for anything conversational. When you go to production, you change the model name and nothing else.

To use your own voice, the docs say an instant clone takes one recording (it uses the first three minutes) and is usually ready in seconds. Instant clones are free while the preview lasts. Professional clones need 20 or more minutes of audio, use a paid clone slot, and only work on the standard heygen-voice-1, not Turbo.

A good first test for a business: have it read your voicemail greeting, a phone number, a street address and an appointment time. Numbers and codes are the category where Turbo scored best, and a misread phone number is the mistake a caller notices first.

Our read

HeyGen also says 64% of its users preferred Turbo over ElevenLabs in blind tests. That's the company's own survey, so give the Artificial Analysis arena more weight. Even there, Turbo leads its predecessor by only 6 Elo points, so we'd call the two a tie on sound and pick on price. In late September we wrote that Eleven v4 topped the voice rankings at 5x Gemini's price, and since then Voxtral 3 and now Turbo have pushed the cost of good speech down hard. If you're building a phone agent, also look at Eleven Realtime, which handles the whole conversation and not only the reading aloud.

For a small business, a natural-sounding voice is no longer the expensive part of an automated phone line, a reminder call or a training video. The work that's left is deciding what the voice should say and connecting it to your calendar and CRM. If you want help figuring out which of your calls or follow-ups make sense to hand off, New Face Design's free process audit is a good place to start.