Alibaba HappyHorse 1.1 is a video-generation edition with separately named text-to-video, first-frame image-to-video and reference-to-video profiles. The manufacturer emphasizes improved visual detail, movement, character identity across clips, text stability and audio-visual synchronization. The result is a short video with sound, useful for a product shot, a portrait animation or a small character scene. Select the input mode from what should define the scene: a written concept, an exact opening image or several visual references.
For text-to-video, describe the subject, location, action, framing, camera movement and sound intent. For a short action scene, show a cyclist stopping beside a riverside café while the camera moves gently toward the parked bicycle. Keep the action appropriate to a short clip and describe the final moment. The maker’s current text guide permits up to 5,000 non-Chinese or 2,500 Chinese characters and truncates excess text. Organize the brief so its essential subject and action are explicit rather than buried in repeated adjectives.
First-frame image-to-video uses exactly one image as the opening. Prepare the subject, composition and intended crop in that still, then describe what should move or how the camera should change. The documented image formats are JPEG/JPG, PNG and WEBP, up to 20 MB, with both dimensions at least 300 pixels and an aspect ratio between 1:2.5 and 2.5:1. The output approximates the selected resolution tier while retaining the source image’s proportions. A portrait result therefore starts with a suitable portrait composition rather than an unrelated aspect-ratio instruction.
Reference-to-video accepts one to nine images. The prompt uses [Image 1], [Image 2] and subsequent positions to identify them in the input order. State the role of each reference and name the relevant object or person, so a scene image is not mistaken for a character source. The reference guide accepts JPEG/JPG, PNG or WEBP up to 20 MB, requires the shortest side to be at least 400 pixels and recommends clear references at 720p or higher. Avoid a heavily compressed reference whose identifying detail is already missing.
For a scene with several references, place the person from one image in the room from another while retaining a coat or product supplied by a third. Describe the interaction and camera in the same brief. The 1.1 reference profile highlights stability of subjects, scene style and cinematic motion; inspect the actual result for the intended identity, product markings and spatial placement. Several images give the brief more information, but the review still needs to check which features carried into the video.
The maker lists 480p, 720p and 1080p output for the 1.1 generation modes, with whole-second duration from three to fifteen and MP4 at 24 fps. The API response describes H.264 encoding. Text and reference generation offer shapes including 16:9, 9:16, 1:1, 4:3, 3:4, 4:5, 5:4, 9:21 and 21:9; first-frame generation follows the input image. Choose the shape for the destination and leave room for later captions or layout elements. Inspect the returned video’s actual framing at the intended display size.
The model family’s output includes sound. Give the performance or atmosphere a clear role in the brief, and review any dialogue, effects or background track with the picture. Keep spoken material short enough for the action and pauses rather than compressing a long script into one short shot. Check lip movement, speaker attribution and the actual wording when speech is present. Prompt-language acceptance does not itself define a guaranteed pronunciation or a separate voice-reference input.
The documented maker workflow is asynchronous: a task identifier tracks processing, and a successful result supplies the video location. The result link and task queries are documented with a twenty-four-hour validity period. Retain the completed clip, references and prompt as the revision record. The guides also document seed and watermark controls. A fixed seed can aid comparison, but the maker explicitly says it does not guarantee identical results, so judge the returned artifact rather than assuming one setting fixes every frame.
HappyHorse 1.0 and 1.1 are separate editions. The 1.1 output table adds 480p to the 720p and 1080p options of 1.0, and the 1.1 model pages describe upgrades in detail, motion and consistency. The separately named HappyHorse 1.0 video-edit profile has its own inputs and limits and is not an editing mode to silently assign to 1.1. Review the complete 1.1 clip for identity, motion, framing, text and sound, then make a specific correction while retaining the successful parts of the brief.
", "aboutText": "Alibaba HappyHorse 1.1 is a video-generation edition with separately named text-to-video, first-frame image-to-video and reference-to-video profiles. The manufacturer emphasizes improved visual detail, movement, character identity across clips, text stability and audio-visual synchronization. The result is a short video with sound, useful for a product shot, a portrait animation or a small character scene. Select the input mode from what should define the scene: a written concept, an exact opening image or several visual references.\n\nFor text-to-video, describe the subject, location, action, framing, camera movement and sound intent. For a short action scene, show a cyclist stopping beside a riverside café while the camera moves gently toward the parked bicycle. Keep the action appropriate to a short clip and describe the final moment. The maker’s current text guide permits up to 5,000 non-Chinese or 2,500 Chinese characters and truncates excess text. Organize the brief so its essential subject and action are explicit rather than buried in repeated adjectives.\n\nFirst-frame image-to-video uses exactly one image as the opening. Prepare the subject, composition and intended crop in that still, then describe what should move or how the camera should change. The documented image formats are JPEG/JPG, PNG and WEBP, up to 20 MB, with both dimensions at least 300 pixels and an aspect ratio between 1:2.5 and 2.5:1. The output approximates the selected resolution tier while retaining the source image’s proportions. A portrait result therefore starts with a suitable portrait composition rather than an unrelated aspect-ratio instruction.\n\nReference-to-video accepts one to nine images. The prompt uses [Image 1], [Image 2] and subsequent positions to identify them in the input order. State the role of each reference and name the relevant object or person, so a scene image is not mistaken for a character source. The reference guide accepts JPEG/JPG, PNG or WEBP up to 20 MB, requires the shortest side to be at least 400 pixels and recommends clear references at 720p or higher. Avoid a heavily compressed reference whose identifying detail is already missing.\n\nFor a scene with several references, place the person from one image in the room from another while retaining a coat or product supplied by a third. Describe the interaction and camera in the same brief. The 1.1 reference profile highlights stability of subjects, scene style and cinematic motion; inspect the actual result for the intended identity, product markings and spatial placement. Several images give the brief more information, but the review still needs to check which features carried into the video.\n\nThe maker lists 480p, 720p and 1080p output for the 1.1 generation modes, with whole-second duration from three to fifteen and MP4 at 24 fps. The API response describes H.264 encoding. Text and reference generation offer shapes including 16:9, 9:16, 1:1, 4:3, 3:4, 4:5, 5:4, 9:21 and 21:9; first-frame generation follows the input image. Choose the shape for the destination and leave room for later captions or layout elements. Inspect the returned video’s actual framing at the intended display size.\n\nThe model family’s output includes sound. Give the performance or atmosphere a clear role in the brief, and review any dialogue, effects or background track with the picture. Keep spoken material short enough for the action and pauses rather than compressing a long script into one short shot. Check lip movement, speaker attribution and the actual wording when speech is present. Prompt-language acceptance does not itself define a guaranteed pronunciation or a separate voice-reference input.\n\nThe documented maker workflow is asynchronous: a task identifier tracks processing, and a successful result supplies the video location. The result link and task queries are documented with a twenty-four-hour validity period. Retain the completed clip, references and prompt as the revision record. The guides also document seed and watermark controls. A fixed seed can aid comparison, but the maker explicitly says it does not guarantee identical results, so judge the returned artifact rather than assuming one setting fixes every frame.\n\nHappyHorse 1.0 and 1.1 are separate editions. The 1.1 output table adds 480p to the 720p and 1080p options of 1.0, and the 1.1 model pages describe upgrades in detail, motion and consistency. The separately named HappyHorse 1.0 video-edit profile has its own inputs and limits and is not an editing mode to silently assign to 1.1. Review the complete 1.1 clip for identity, motion, framing, text and sound, then make a specific correction while retaining the successful parts of the brief.", "editorialOverlay": { "copy_fields": null, "copy_sha256": null, "copy_source": null, "html_sha256": "1cc3e78a5a841e4936b8813ec9c310e6d3d4a676a8332d7a4055716cde8fc566", "module_sha256": "ee9dc1c50b1e6a8fb7fa062030fcb4646b9aaaeb08be513b723a5a92a7290f06", "raw_model_unchanged": true, "schema": "ailoot-native-model-editorial-v1" }, "locale": "en", "methods": [ { "html": "alibaba/happyhorse-1.1Alibaba HappyHorse 1.1 generates short videos with sound from a text brief, one first-frame image or up to nine reference images. Its maker profiles emphasize improved motion, skin detail and cross-clip character consistency. Output is 3–15-second MP4 at 24 fps, with 480p, 720p or 1080p options.
This description covers the model’s capabilities. Available in Ailoot inputs, outputs and options are listed for the selected method below.
Alibaba HappyHorse 1.1 is a video-generation edition with separately named text-to-video, first-frame image-to-video and reference-to-video profiles. The manufacturer emphasizes improved visual detail, movement, character identity across clips, text stability and audio-visual synchronization. The result is a short video with sound, useful for a product shot, a portrait animation or a small character scene. Select the input mode from what should define the scene: a written concept, an exact opening image or several visual references.
For text-to-video, describe the subject, location, action, framing, camera movement and sound intent. For a short action scene, show a cyclist stopping beside a riverside café while the camera moves gently toward the parked bicycle. Keep the action appropriate to a short clip and describe the final moment. The maker’s current text guide permits up to 5,000 non-Chinese or 2,500 Chinese characters and truncates excess text. Organize the brief so its essential subject and action are explicit rather than buried in repeated adjectives.
First-frame image-to-video uses exactly one image as the opening. Prepare the subject, composition and intended crop in that still, then describe what should move or how the camera should change. The documented image formats are JPEG/JPG, PNG and WEBP, up to 20 MB, with both dimensions at least 300 pixels and an aspect ratio between 1:2.5 and 2.5:1. The output approximates the selected resolution tier while retaining the source image’s proportions. A portrait result therefore starts with a suitable portrait composition rather than an unrelated aspect-ratio instruction.
Reference-to-video accepts one to nine images. The prompt uses [Image 1], [Image 2] and subsequent positions to identify them in the input order. State the role of each reference and name the relevant object or person, so a scene image is not mistaken for a character source. The reference guide accepts JPEG/JPG, PNG or WEBP up to 20 MB, requires the shortest side to be at least 400 pixels and recommends clear references at 720p or higher. Avoid a heavily compressed reference whose identifying detail is already missing.
For a scene with several references, place the person from one image in the room from another while retaining a coat or product supplied by a third. Describe the interaction and camera in the same brief. The 1.1 reference profile highlights stability of subjects, scene style and cinematic motion; inspect the actual result for the intended identity, product markings and spatial placement. Several images give the brief more information, but the review still needs to check which features carried into the video.
The maker lists 480p, 720p and 1080p output for the 1.1 generation modes, with whole-second duration from three to fifteen and MP4 at 24 fps. The API response describes H.264 encoding. Text and reference generation offer shapes including 16:9, 9:16, 1:1, 4:3, 3:4, 4:5, 5:4, 9:21 and 21:9; first-frame generation follows the input image. Choose the shape for the destination and leave room for later captions or layout elements. Inspect the returned video’s actual framing at the intended display size.
The model family’s output includes sound. Give the performance or atmosphere a clear role in the brief, and review any dialogue, effects or background track with the picture. Keep spoken material short enough for the action and pauses rather than compressing a long script into one short shot. Check lip movement, speaker attribution and the actual wording when speech is present. Prompt-language acceptance does not itself define a guaranteed pronunciation or a separate voice-reference input.
The documented maker workflow is asynchronous: a task identifier tracks processing, and a successful result supplies the video location. The result link and task queries are documented with a twenty-four-hour validity period. Retain the completed clip, references and prompt as the revision record. The guides also document seed and watermark controls. A fixed seed can aid comparison, but the maker explicitly says it does not guarantee identical results, so judge the returned artifact rather than assuming one setting fixes every frame.
HappyHorse 1.0 and 1.1 are separate editions. The 1.1 output table adds 480p to the 720p and 1080p options of 1.0, and the 1.1 model pages describe upgrades in detail, motion and consistency. The separately named HappyHorse 1.0 video-edit profile has its own inputs and limits and is not an editing mode to silently assign to 1.1. Review the complete 1.1 clip for identity, motion, framing, text and sound, then make a specific correction while retaining the successful parts of the brief.
POST /v1/videostextvideo| Field | Type | Requirement | Values & limits | Default |
|---|---|---|---|---|
Fieldmodel | Typestring | RequirementRequired | Values & limits alibaba/happyhorse-1.1 | Default— |
Fieldprompt | Typestring | RequirementRequired | Values & limits Minimum length: 1 Maximum bytes: 10,000 Unit: UTF8bytes | Default— |
Fieldduration | Typeinteger | RequirementRequired | Values & limits Minimum: 3 Maximum: 15 Unit: seconds | Default— |
Fieldresolution | Typestring | RequirementRequired | Values & limits 720p | Default— |
Fieldaspect_ratio | Typestring | RequirementRequired | Values & limits 16:9 | Default— |
Text to video · Required: prompt
| Parameter | Contract |
|---|---|
Parametercontent_type | Contractapplication/json |
Parameterrequired | Contractmodel, prompt, duration, resolution, aspect_ratio |
Parameterother_options | Contractfalse |
Parameterresolution | Contract720p |
Parameterprompt_max_bytes | Contract10000 |
Parameterduration_integer_range | Contract3, 15 |
Parameteraspect_ratio | Contract16:9 |
Parametergenerate_audio | Contractmust be omitted (not false); AAC included |
Parametermaterial_types | Contracttext |
Parameteradditional_properties | Contractfalse |
Method aliases: /v1/videos/generations, /videos, /videos/generations
Lifecycle: submit_once_poll_original
GET /v1/videos/{request_id}
GET /v1/videos/generations/{request_id}
GET /v1/videos/{request_id}/content
GET /v1/videos/generations/{request_id}/content
Create once and save returned ID.
Poll original status; retain every completed asset and duration.
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 conditionsresolution: 1080p | Charge Output · Video price · Estimate for comparison only $0.2556 per video second |
| ConfigurationVideo price | Rate schedule— | Forecast quantity— | Billing unitper video second | Billing conditionsresolution: 720p | Charge Output · Video price · Estimate for comparison only $0.1976 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 $0.5928 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 →
Model-level reference schedules do not change method selection, quotes or charges. Active charges use the selected Ailoot method rates.
Match each charge to the conditions shown. Use one rate for each set of alternative configurations.
| Configuration | Billing conditions | Charge |
|---|---|---|
| ConfigurationVideo price | Billing conditionsresolution: 720p | Charge Output · Video price $0.1976 per video second Published price reference for the configuration shown. Final charges follow the actual billed configuration. |
| ConfigurationVideo price | Billing conditionsresolution: 1080p | Charge Output · Video price $0.2556 per video second Published price reference for the configuration shown. Final charges follow the actual billed configuration. |
set -eu\n: "${AILOOT_API_KEY:?Set your Ailoot API key}"\npython3 - <<'PY_REQUEST'\nimport json,os\nfrom pathlib import Path\np=json.loads('{"model": "alibaba/happyhorse-1.1", "prompt": "A cat in a spacesuit walking on the Moon", "duration": 3, "resolution": "720p", "aspect_ratio": "16:9"}')\n\nPath("request.json").write_text(json.dumps(p,ensure_ascii=False))\nPY_REQUEST\ncurl --fail-with-body --silent --show-error https://api.ailoot.ai/v1/videos \\\n -H "Authorization: Bearer $AILOOT_API_KEY" -H 'Content-Type: application/json' \\\n --data-binary @request.json -D response-headers.txt -o result.json\nmp4
GET /v1/videos/{request_id}GET /v1/videos/generations/{request_id}GET /v1/videos/{request_id}/contentGET /v1/videos/generations/{request_id}/contentset -eu\n: "${AILOOT_API_KEY:?Set your Ailoot API key}"\npython3 - <<'PY_STATUS_URL'\nimport json\nfrom pathlib import Path\nfrom urllib.parse import quote\nr=json.loads(Path('result.json').read_text())\nif not isinstance(r.get('id'),str) or not r['id']:raise SystemExit('Keep the original response and headers for reconciliation; no returned ID')\nPath('status-url.txt').write_text('https://api.ailoot.ai/v1/videos/'+quote(r['id'],safe=''))\nPY_STATUS_URL\ncurl --fail-with-body --silent --show-error "$(cat status-url.txt)" \\\n -H "Authorization: Bearer $AILOOT_API_KEY" -D status-headers.txt -o status.json\npython3 - <<'PY_RESULT'\nimport json\nfrom pathlib import Path\nr=json.loads(Path('status.json').read_text())\nprint(json.dumps(r,ensure_ascii=False,indent=2))\nprint('Keep the complete JSON and response headers; do not resubmit an unknown creation.')\nPY_RESULT\nset -eu\n: "${AILOOT_API_KEY:?Set your Ailoot API key}"\nexport AILOOT_API_KEY\npython3 - <<'PY_ASSETS'\nimport json,os,subprocess\nfrom pathlib import Path\nfrom urllib.parse import urljoin,urlsplit\nr=json.loads(Path('status.json').read_text())\nif r.get('status')!='completed':raise SystemExit('Keep the original ID and read its status; never recreate an unknown job')\nif not isinstance(r.get('data'),list) or not r['data']:raise SystemExit('Keep the complete original status; no completed outputs supplied')\nurls=[]\nfor i,u in enumerate(r.get('data',[])):\n if not isinstance(u,str):raise SystemExit('Keep the original response; unexpected asset shape')\n urls.append((str(i),u))\nfor label,path in urls:\n u=urljoin('https://api.ailoot.ai',path);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 returned Ailoot owned content URL')\n target=Path('owned-'+label+'.bin')\n subprocess.run(['curl','--fail-with-body','--silent','--show-error',u,'-H','Authorization: Bearer '+os.environ['AILOOT_API_KEY'],'-D',str(target)+'.headers','-o',str(target)],check=True)\nprint('All returned owned URLs saved; complete metadata, timing and usage remain in the JSON.')\nPY_ASSETS\nSave every completed video or returned file; retrieve every actual returned asset URL.
Completed data[] contains all owned signed content URLs. Fetch each returned URL, or use the content endpoint with index set to an actual output index. A failed original job may return charge_pending_check; check that original job rather than creating it again.
Read existing original job; never resubmit unknown send.
Maximum outputs: 4.
| Field | Type | Meaning |
|---|---|---|
Fieldid | Typestring | Meaning— |
Fieldobject | Typestring | Meaning— |
Fieldmodel | Typestring | Meaning— |
Fieldstatus | Typestring | Meaning— |
Fielddata[] | Typesigned content URL | Meaning— |
Fieldcharge_pending_check | Typeboolean | Meaning— |
Fielderror.code | Typestring | Meaning— |
Fielderror.message | Typestring | Meaning— |
Fieldoperation_id | Typestring | MeaningHTTP503 result_pending_check original status/content response: same original public video ID as id. |
Fielderror.type | Typestring | Meaningserver_error on HTTP503 result_pending_check. |
| Response | Condition | Next step |
|---|---|---|
Response400 · invalid_request_error | ConditionReading the raw request body fails before parsing the video profile | Next stepPreserve the returned response and correct the unreadable request body. |
Response422 · price_unavailable | ConditionParsing the exact video profile fails, or the complete parsed profile has no ready exact price card | Next stepPreserve the response; use the complete exact admitted profile and its published card. This branch runs before a provider send. |
Response409 · video_not_ready | ConditionOriginal result content is not yet available | Next stepPoll the original returned ID and respect Retry-After. |
Responseserver_error | ConditionThe returned response says the operation or result requires a check | Next stepKeep the original returned operation/task ID, request headers and response; continue supported reads of that original operation. Do not resubmit the creation request. |
Response503 · result_pending_check | ConditionGET original status/content: task is not published, original job is not failed/cancelled/expired, and local state is unknown or original job status is completed. | Next stepKeep original id and operation_id (same public video ID), response and headers; read only the original status/content. Never submit a new generation request. |
Ailoot Ultra alibaba/happyhorse-1.1
/v1/videos/generations, /videos, /videos/generations