Here's what's new in the HeyGen API this month. The HeyGen Video API is live, you can now create templates through the API, and ElevenLabs v4 is available for speech. Professional Voice Clone and Look Packs have also arrived, with simpler scene editing in Video Agent. Each update links to a guide to help you get started. Please give it a read, and let us know what you build.

Highlights

HeyGen Video 1.0 API

Generate short videos from a prompt, an image or reference media.

Get started

HeyGen Voice: Professional Voice Clone

Train a studio-quality voice clone on HeyGen's own voice model, then generate speech with it, either as one finished file or streamed.

Get started

ElevenLabs v4 for speech

ElevenLabs v4 is available for speech, and you can choose an ElevenLabs model for each request.

Get started

Burst concurrency

Run more jobs at once during busy periods, without changing your plan.

Get started

Translation rules on brand glossaries

Set terms that should never be translated, and terms that should always be translated a certain way.

Get started

Look Packs over the API

Give an avatar you already have a new set of looks in one call. You can also list the Look Packs your workspace can use, including your enterprise packs.

Get started

Edit specific scenes in Video Agent

Change only the scenes you name and keep the rest of the video as it is.

Get started

Create templates through the API

Turn any video into a template and generate a personalized version for each viewer.

Get started

Folders

Create folders and send your videos and translations straight into them. They appear in the HeyGen app.

Get started

Also new

  • Check your API key. GET /v3/api_keys/self shows a key's permissions and expiry date. Every endpoint in the reference also lists the permissions it needs.

  • Manage proofread sessions. List and delete Video Translate proofreads.

  • Video Agent sharing. Choose who can open a session when you create it: just you, your workspace, or specific people.

  • Video quality checks support more criteria and up to 2,000 comparisons per batch.

  • Bigger uploads. Audio files can be up to 1.5 GB.

Good to know

  • Professional Voice Clone endpoints now live under /v3/models/audio/voices. See the guide

  • Speech engine. When you don't choose an engine on /v3/voices/speech, the voice's default engine is used. To keep using Starfish, set engine: "starfish".

  • Brand policy. If your workspace locks its brand, Video Agent uses the workspace's default brand kit automatically.