Avatar IV is HeyGen’s avatar engine for turning a still image into an expressive talking video. Its diffusion-inspired audio-to-expression design uses vocal tone, rhythm and emotion to guide facial movement, head tilts, pauses and micro-expressions. The purpose is a presenter or character delivering spoken content with lip sync and gestures, rather than a general scene generator that creates a whole setting and soundtrack from a cinematic prompt.
The manufacturer describes support for lifelike and stylized characters, including angled or profile images, anime-style faces and pets. Photo-based characters work best when their eyes, mouth and facial structure are easy to identify. HeyGen recommends a clear, well-lit image with a natural posture and visible hands when gestures matter. A very small face, hidden lips, heavy masks, a pronounced beak or highly exaggerated proportions can weaken facial movement and lip sync. Keep the source framing suited to the intended performance.
The producer provides two speech-input paths. A written script is paired with a selected voice; a prerecorded audio track supplies the voice and timing directly. Use the recorded-audio path when the delivery is already established, and the script path when the speech needs to be synthesized. These inputs are alternatives in the documented interface. Prepare the final wording, names and pronunciation before animation, and preserve the audio or script with the returned video so the spoken result can be checked.
HeyGen’s Avatar IV model profile covers studio avatars, Digital Twins, Photo Avatars, direct images and prompt-created subjects. A reusable Photo Avatar provides a stable subject for later videos, while a direct-image workflow starts from the supplied picture. The producer’s Photo Avatar profile offers free-text motion instructions and expressiveness choices of high, medium or low, with low as its default. Retain the chosen avatar look and engine when making a series so a changed source image or rendering model does not become an unnoticed change in the presenter.
For motion direction, describe a visible body action and its emotional quality. “Nods gently with a warm smile” gives one gesture and one expression; a brief greeting can use a friendly wave. The maker recommends one main gesture, optionally paired with a facial expression, and no more than two short clauses for a sequence. Avatar IV handles pacing from the performance, so the Studio guide does not require a second-by-second gesture schedule. Specific, concise wording is easier to inspect than several competing actions in one prompt.
The producer’s audio-to-video profile renders up to thirty minutes of supplied audio per request, and uploaded MP3 or WAV assets have a 32 MB limit. For a longer lesson or recording, divide the material at coherent speech boundaries and retain the original ordering. The final video’s spoken timing should follow the selected audio rather than an unrelated clip-duration estimate.
The documented producer output settings include 720p or 1080p and shapes of 16:9, 9:16, 4:5, 5:4 or 1:1, with an automatic option that follows the source proportions. A cover fit can crop the subject; contain fits the whole subject and may leave surrounding background visible. Its audio-to-video workflow returns MP4 by default. For video avatars, transparent WebM output requires matting support, and the producer applies background removal for that output rather than combining it with a new background. Inspect the actual framing and file before placing the presenter in a larger composition.
A training introduction can use a readable presenter image, a concise explanation and one restrained gesture at the key instruction. A product explainer can retain the same avatar look while changing the spoken script for different features. The producer also documents optional subtitle files and burned captions; review their wording and timing alongside the speech. Rendering is asynchronous in the maker interface, so a creation response and the completed video are different artifacts. Keep the finished recording and its completion state with the task’s result.
Avatar IV remains distinct from Avatar V’s reference-driven motion and its separate Studio preset controls. It is also distinct from HeyGen Video’s general-purpose scene generation and from an interactive live-avatar conversation. Review the rendered mouth movement, visible hands, facial expression, pauses and spoken wording against the source material and intended tone. Image, pose, lighting and motion instructions affect the result; a photo-based animation should be checked as a generated performance, with the chosen inputs retained for later revisions.
", "aboutText": "Avatar IV is HeyGen’s avatar engine for turning a still image into an expressive talking video. Its diffusion-inspired audio-to-expression design uses vocal tone, rhythm and emotion to guide facial movement, head tilts, pauses and micro-expressions. The purpose is a presenter or character delivering spoken content with lip sync and gestures, rather than a general scene generator that creates a whole setting and soundtrack from a cinematic prompt.\n\nThe manufacturer describes support for lifelike and stylized characters, including angled or profile images, anime-style faces and pets. Photo-based characters work best when their eyes, mouth and facial structure are easy to identify. HeyGen recommends a clear, well-lit image with a natural posture and visible hands when gestures matter. A very small face, hidden lips, heavy masks, a pronounced beak or highly exaggerated proportions can weaken facial movement and lip sync. Keep the source framing suited to the intended performance.\n\nThe producer provides two speech-input paths. A written script is paired with a selected voice; a prerecorded audio track supplies the voice and timing directly. Use the recorded-audio path when the delivery is already established, and the script path when the speech needs to be synthesized. These inputs are alternatives in the documented interface. Prepare the final wording, names and pronunciation before animation, and preserve the audio or script with the returned video so the spoken result can be checked.\n\nHeyGen’s Avatar IV model profile covers studio avatars, Digital Twins, Photo Avatars, direct images and prompt-created subjects. A reusable Photo Avatar provides a stable subject for later videos, while a direct-image workflow starts from the supplied picture. The producer’s Photo Avatar profile offers free-text motion instructions and expressiveness choices of high, medium or low, with low as its default. Retain the chosen avatar look and engine when making a series so a changed source image or rendering model does not become an unnoticed change in the presenter.\n\nFor motion direction, describe a visible body action and its emotional quality. “Nods gently with a warm smile” gives one gesture and one expression; a brief greeting can use a friendly wave. The maker recommends one main gesture, optionally paired with a facial expression, and no more than two short clauses for a sequence. Avatar IV handles pacing from the performance, so the Studio guide does not require a second-by-second gesture schedule. Specific, concise wording is easier to inspect than several competing actions in one prompt.\n\nThe producer’s audio-to-video profile renders up to thirty minutes of supplied audio per request, and uploaded MP3 or WAV assets have a 32 MB limit. For a longer lesson or recording, divide the material at coherent speech boundaries and retain the original ordering. The final video’s spoken timing should follow the selected audio rather than an unrelated clip-duration estimate.\n\nThe documented producer output settings include 720p or 1080p and shapes of 16:9, 9:16, 4:5, 5:4 or 1:1, with an automatic option that follows the source proportions. A cover fit can crop the subject; contain fits the whole subject and may leave surrounding background visible. Its audio-to-video workflow returns MP4 by default. For video avatars, transparent WebM output requires matting support, and the producer applies background removal for that output rather than combining it with a new background. Inspect the actual framing and file before placing the presenter in a larger composition.\n\nA training introduction can use a readable presenter image, a concise explanation and one restrained gesture at the key instruction. A product explainer can retain the same avatar look while changing the spoken script for different features. The producer also documents optional subtitle files and burned captions; review their wording and timing alongside the speech. Rendering is asynchronous in the maker interface, so a creation response and the completed video are different artifacts. Keep the finished recording and its completion state with the task’s result.\n\nAvatar IV remains distinct from Avatar V’s reference-driven motion and its separate Studio preset controls. It is also distinct from HeyGen Video’s general-purpose scene generation and from an interactive live-avatar conversation. Review the rendered mouth movement, visible hands, facial expression, pauses and spoken wording against the source material and intended tone. Image, pose, lighting and motion instructions affect the result; a photo-based animation should be checked as a generated performance, with the chosen inputs retained for later revisions.", "editorialOverlay": { "copy_fields": null, "copy_sha256": null, "copy_source": null, "html_sha256": "e2a938a8ced2f7ff96c545836632ce8f86940fc2406a7daf4833a71c67b84452", "module_sha256": "ee9dc1c50b1e6a8fb7fa062030fcb4646b9aaaeb08be513b723a5a92a7290f06", "raw_model_unchanged": true, "schema": "ailoot-native-model-editorial-v1" }, "locale": "en", "methods": [ { "html": "heygen/avatar-ivAvatar IV animates a photo into a lip-synced talking video, with facial expression and gesture timing guided by the voice. HeyGen documents script or recorded-audio workflows and Photo Avatar motion controls for lifelike or stylized presenters.
This description covers the model’s capabilities. Available in Ailoot inputs, outputs and options are listed for the selected method below.
Avatar IV is HeyGen’s avatar engine for turning a still image into an expressive talking video. Its diffusion-inspired audio-to-expression design uses vocal tone, rhythm and emotion to guide facial movement, head tilts, pauses and micro-expressions. The purpose is a presenter or character delivering spoken content with lip sync and gestures, rather than a general scene generator that creates a whole setting and soundtrack from a cinematic prompt.
The manufacturer describes support for lifelike and stylized characters, including angled or profile images, anime-style faces and pets. Photo-based characters work best when their eyes, mouth and facial structure are easy to identify. HeyGen recommends a clear, well-lit image with a natural posture and visible hands when gestures matter. A very small face, hidden lips, heavy masks, a pronounced beak or highly exaggerated proportions can weaken facial movement and lip sync. Keep the source framing suited to the intended performance.
The producer provides two speech-input paths. A written script is paired with a selected voice; a prerecorded audio track supplies the voice and timing directly. Use the recorded-audio path when the delivery is already established, and the script path when the speech needs to be synthesized. These inputs are alternatives in the documented interface. Prepare the final wording, names and pronunciation before animation, and preserve the audio or script with the returned video so the spoken result can be checked.
HeyGen’s Avatar IV model profile covers studio avatars, Digital Twins, Photo Avatars, direct images and prompt-created subjects. A reusable Photo Avatar provides a stable subject for later videos, while a direct-image workflow starts from the supplied picture. The producer’s Photo Avatar profile offers free-text motion instructions and expressiveness choices of high, medium or low, with low as its default. Retain the chosen avatar look and engine when making a series so a changed source image or rendering model does not become an unnoticed change in the presenter.
For motion direction, describe a visible body action and its emotional quality. “Nods gently with a warm smile” gives one gesture and one expression; a brief greeting can use a friendly wave. The maker recommends one main gesture, optionally paired with a facial expression, and no more than two short clauses for a sequence. Avatar IV handles pacing from the performance, so the Studio guide does not require a second-by-second gesture schedule. Specific, concise wording is easier to inspect than several competing actions in one prompt.
The producer’s audio-to-video profile renders up to thirty minutes of supplied audio per request, and uploaded MP3 or WAV assets have a 32 MB limit. For a longer lesson or recording, divide the material at coherent speech boundaries and retain the original ordering. The final video’s spoken timing should follow the selected audio rather than an unrelated clip-duration estimate.
The documented producer output settings include 720p or 1080p and shapes of 16:9, 9:16, 4:5, 5:4 or 1:1, with an automatic option that follows the source proportions. A cover fit can crop the subject; contain fits the whole subject and may leave surrounding background visible. Its audio-to-video workflow returns MP4 by default. For video avatars, transparent WebM output requires matting support, and the producer applies background removal for that output rather than combining it with a new background. Inspect the actual framing and file before placing the presenter in a larger composition.
A training introduction can use a readable presenter image, a concise explanation and one restrained gesture at the key instruction. A product explainer can retain the same avatar look while changing the spoken script for different features. The producer also documents optional subtitle files and burned captions; review their wording and timing alongside the speech. Rendering is asynchronous in the maker interface, so a creation response and the completed video are different artifacts. Keep the finished recording and its completion state with the task’s result.
Avatar IV remains distinct from Avatar V’s reference-driven motion and its separate Studio preset controls. It is also distinct from HeyGen Video’s general-purpose scene generation and from an interactive live-avatar conversation. Review the rendered mouth movement, visible hands, facial expression, pauses and spoken wording against the source material and intended tone. Image, pose, lighting and motion instructions affect the result; a photo-based animation should be checked as a generated performance, with the chosen inputs retained for later revisions.
POST /v1/videostext, image, audio, videovideo| Field | Type | Requirement | Values & limits | Default |
|---|---|---|---|---|
Fieldmodel | Typestring | RequirementRequired | Values & limits heygen/avatar-iv | Default— |
Fieldinput_references | Typearray | RequirementRequired | Values & limits Minimum items: 2 Maximum items: 2 | Default— |
Fieldprompt | Typestring | RequirementOptional | Values & limits | Default— |
| Parameter | Contract |
|---|---|
Parametercontent_type | Contractapplication/json |
Parameterrequired | Contractmodel, input_references |
Parametermaterial_types | Contracttext, image, audio, video |
Parametermaterial_conditions | ContractUse actual completed owned media asset content URLs with their returned signatures: one video source for BFL, or one image and one audio source for avatar. The Source owner/type/byte witness is required., Supply exactly one PNG/JPEG image and one WAV/MP3 audio asset. The audio is at most32000000 bytes, with duration greater than0 and at most1800 seconds. |
Lifecycle: asynchronous_job_then_owned_outputs
GET /v1/videos/{request_id} · actual returned id or polling_url
GET /v1/videos/{request_id}/content
The displayed rates estimate the cost before generation. The final charge is calculated from the cost recorded for the original generation. Measured output duration and forecast quantities are not the final billable quantity.
Tariff references do not add request options. Supported parameters are listed for the selected method.
Match each charge to the conditions shown. Use one rate for each set of alternative configurations.
| Configuration | Rate schedule | Forecast quantity | Billing unit | Billing conditions | Charge |
|---|---|---|---|---|---|
| ConfigurationBase token rates | Rate scheduleBase rate | Forecast quantity— | Billing unit— | Billing conditions— | Charge Input · Input tokens · Estimate for comparison only — / 1K tokens Output · Output tokens · Estimate for comparison only — / 1K tokens Input · Cache read · Estimate for comparison only — / 1K tokens Input · Cache write · Estimate for comparison only — / 1K tokens Input · Cache write, 1 hour · Estimate for comparison only — / 1K tokens |
| ConfigurationVideo price | Rate schedule— | Forecast quantity— | Billing unitper video second | Billing conditionsNo additional condition is declared in this pricing key. | Charge Output · Video price · Estimate for comparison only $0.1 per video second |
| ConfigurationEstimated generation cost | Rate schedule— | Forecast quantityone generation estimate | Billing unit— | Billing conditions— | Charge Output · Estimated generation cost · Estimate for comparison only $0.1 estimate before generation This estimate is not an additional charge or a fixed rate for the completed generation. |
Ailoot reserves an estimated amount before a request. After the result, it settles the final charge from measured usage. Streaming settlement uses the final usage record. How reservation and settlement work →
set -eu\n: "${AILOOT_API_KEY:?Set your Ailoot API key}"\n: "${AILOOT_INPUT_REFERENCES_JSON:?Set the actual documented input value}"\nexport AILOOT_INPUT_REFERENCES_JSON\npython3 - <<'PY_REQUEST'\nimport json,os\nfrom pathlib import Path\nrequest={'model': 'heygen/avatar-iv', 'input_references': '${ACTUAL_OWNED_INPUT_REFERENCES}'}\nrequest['input_references']=json.loads(os.environ['AILOOT_INPUT_REFERENCES_JSON'])\nPath('request.json').write_text(json.dumps(request,ensure_ascii=False),encoding='utf-8')\nPY_REQUEST\ncurl --fail-with-body --silent --show-error https://api.ailoot.ai/v1/videos \\\n -H "Authorization: Bearer $AILOOT_API_KEY" -H 'Content-Type: application/json' \\\n --data-binary @request.json -D response-headers.txt -o response.bin\n# Retain the complete response and headers, every actual output, original returned identifiers and all usage fields. Recover the original operation without submitting creation again.\nGET /v1/videos/{request_id}GET /v1/videos/{request_id}/contentset -eu\n: "${AILOOT_API_KEY:?Set your Ailoot API key}"\n: "${AILOOT_STATUS_FILE:?Set the actual completed status JSON file}"\nexport AILOOT_STATUS_FILE\npython3 - <<'PY_REQUEST'\nimport json,os\nfrom pathlib import Path\nfrom urllib.parse import urlsplit,urljoin\nf=Path(os.environ["AILOOT_STATUS_FILE"])\nif not f.is_file():raise SystemExit("Choose the actual completed status file")\nr=json.loads(f.read_text())\nif r.get("status")!="completed" or not isinstance(r.get("data"),list) or not r["data"]:raise SystemExit("Wait for a completed status with asset paths")\nurls=[]\nfor x in r["data"]:\n if not isinstance(x,str) or "\\n" in x or "\\r" in x:raise SystemExit("Invalid asset path")\n u=urljoin("https://api.ailoot.ai",x)\n z=urlsplit(u)\n if z.scheme!="https" or z.netloc!="api.ailoot.ai" or z.username or z.password:raise SystemExit("Use the actual Ailoot asset paths from status")\n urls.append(u)\nPath("asset-urls.txt").write_text("\\n".join(urls)+"\\n")\nPY_REQUEST\nindex=0\nwhile IFS= read -r asset_url; do\n curl --fail-with-body --silent --show-error "$asset_url" \\\n -H "Authorization: Bearer $AILOOT_API_KEY" -o "video-${index}.mp4"\n index=$((index + 1))\ndone < asset-urls.txt\nSave every completed video or returned file; retrieve every actual returned asset URL.
Retain the complete response and headers, every actual output, original returned identifiers and all usage fields. Recover the original operation without submitting creation again.
Keep X-Client-Request-ID, X-Ailoot-Task-Id and actual returned id/polling_url. Continue the original job; do not submit creation again.
| Field | Type |
|---|---|
Fieldid | Type |
Fieldstatus | Type |
Fielddata[] | Type |
Fieldoutputs[] | Type |
Ailoot Ultra heygen/avatar-iv
Use the method paths shown above.