HeyGen tools

HeyGen: Avatar IV is an image-to-video model that animates a single photo into an expressive, lip-synced talking-head video.

Turn a text prompt into a video clip with generated sound (dialogue, effects, ambience, music) using heygen-video-1, HeyGen's model built on MiniMax H3. 5 to 30 seconds at 480p, 768p, 1080p or 2k, in 21:9 to 9:16. Spoken lines in double quotes are voiced; duration auto sizes the clip to the dialogue. Billed per second of output; clips over 15 seconds cost more per second. Use image-to-video to start from a picture, reference-to-video to keep a character, place or voice consistent.

Bring a still image to life: heygen-video-1 (HeyGen's model built on MiniMax H3) starts from your picture and follows the prompt for motion, camera and sound (dialogue in double quotes is voiced). 5 to 30 seconds at 480p, 768p, 1080p or 2k; the output keeps the image's proportions. Same per-second price as text-to-video; clips over 15 seconds cost more per second. Use reference-to-video to combine several images, videos or audio.

Keep characters, products, places, motion or voices consistent: heygen-video-1 (HeyGen's model built on MiniMax H3) generates a clip guided by reference images, videos and audio that the prompt names by position (<Picture 1>, <Video 1>, <Audio 1>). 5 to 15 seconds at 480p, 768p, 1080p or 2k. Costs more per second than text- or image-to-video, plus the reference-video seconds it reads and each reference image after the first two.

Read one voice and its status.

Generate speech from up to 5,000 characters of text with HeyGen Voice 1. Speak with any public voice from /voices or a voice you cloned with /voices/clone; the language is detected from the text or set explicitly. heygen-voice-1 gives the best quality; heygen-voice-1-turbo is several times faster and about 7x cheaper. Returns a 44.1 kHz WAV link, the audio duration and, on request, per-character and per-word timings for captions.