Dža ko lil

H3 Max Reference to Video

Integriruj

Modelosko id minimax/h3-max/reference-to-video
# Run the model; wait up to 20 s for the finished result in the same call
curl -X POST "https://vit.doitong.com/api/v1/run/minimax/h3-max/reference-to-video?wait=20" \
  -H "Authorization: Bearer $DOITONG_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "prompt": "A red fox trots through fresh snow at sunrise, soft golden light, shallow depth of field",
    "prompt_expansion_mode": "balanced"
  }'

# Still "queued" or "running"? Poll with the id from the response:
curl "https://vit.doitong.com/api/v1/requests/REQUEST_ID?wait=20" \
  -H "Authorization: Bearer $DOITONG_API_KEY"

Fajlo sar o misalo: $0.40 400 API krediti tire fajlo si merime angl-o phiripen Phir les kate

Lel API klidin
MiniMax Referenca ko Video Referencaki inputura0.92–15 sdži 1080P #stylized#transform#typography

H3 Max video-generaciǎ kaj astarel sajekh subjektura le referencikane imàginença, videonça thaj audio-kliponça

MiniMax · Video

Tiro prvo rezultato si kate

Ker tele jekh mangipen thaj phir les tire API klidinosa. E rezultatur si kopirime amare arakhipneste, vično vaj 7 dives bilačhes.

Ker jekh mangipen

O kontrolura kerde si katar e schemaske. O kodo thaj e cena menjinen sar tu len menjines.

•trubul 88 / 50000

Video-generation instrukcije referencenca sikavde asset type taj position ande sako reference list, labelenca sar Audio 1, Video 1, Image 1, Image 2, taj avre.

•trubul

Promptosko dorikheripasko zor angle keripe: 'quality' kerel jekh barvaleder prompto maj but sar 30s, taj 'balanced' delpes ande jekh sekunda. Avrirod 'disabled' te na kerel promptosko pherdipe.

resolution standardno 768P

Keripasko rezolucija: 1080P mangela latentno dorikheripe e materialosko kerde native 768P-te, na native keripe ande avrirodli rezolucija.

aspect_ratio standardno adaptive

Aspect ratio vaš o rezultatosko video.

standardno 5

Manglo durata ande sekunda, savo mukel realno video dži 0.7 s maj dur sar manglo.

Lučhe images · reference_video_urls · reference_audio_urls · image · middle_image_url · end_image · middle_frame_time · seed · enable_safety_checker · reference_conditions
reference_image_urls
Čhiv jekh fajl, vaj · dži 9

Subject vaj style reference kipesko URL, sajekh 12-fajlo maximumesa savore video taj audio referencesa. Sikav e kipi ando prompto e listako positionesa: Image 1, Image 2, taj avre labels.

Čhiv jekh fajl, vaj · dži 3

URL adresi le miškipeske vaj aver referentne videonge, prinžarde palal i pozicija ande lista ando prompto pe etiketenca Video 1, Video 2 thaj aver pala lende. O khetano limito le kipenge, videonge thaj audio referencenge si 12 fajlura; sako video klipo trubul te avel 2–15 sekunde, thaj sa khetane na maj but sar 15 sekunde.

Čhiv jekh fajl, vaj · dži 3

De referentno audio sar URL adresi thaj prinžar le klipura ando prompto pe etiketenca Audio 1, Audio 2 thaj aver pala lende. Sako klipo trubul te avel 2–15 sekunde, thaj sasto audio na maj but sar 15 sekunde. Le kipenge, videonge thaj audio referencenge tipura šaj te den pes korkore vaj khetane, dži ko khetano limito 12 fajlura.

image_url
Čhiv jekh fajl, vaj

Opcionalno kipesko URL savo sikavel o exacto anglutno kadro taj kerel o rezultatosko canvas. Reference assets arakhen e subjectosko consistency.

Čhiv jekh fajl, vaj

Opcionalno guidance kipo vaš middle_frame_time, dostupno numa native 480P vaj 768P keripasa taj kana si delime angluno taj agorutno kipo.

end_image_url
Čhiv jekh fajl, vaj

Opcionalno agorutno-kiposko URL savo fixinel o videosko exacto agorutno kadro pe delime kipos.

Middle-image placement time, šaj te lelpe numa khetane middle_image_urlesa taj numa ande pozicije striktno maškar e mangline videoske anglune taj agorutne kadro. Del sekunda save nakhle katar o agor; o value si rounded ko maj pašesko 24 fps kadro.

Bijočaripasko semeno vaš random keripe; te na delpes, jekh randomno avralas.

Reference list delime sar alternativa vaš reference_*_urls, na khetane lenca. E ordering maškar modalities si významno, vaš kodo e sequence arakhelpe savore processingosko.

POST /run/minimax/h3-max/reference-to-video?wait=20 hesab → $0.40
# Run the model; wait up to 20 s for the finished result in the same call
curl -X POST "https://vit.doitong.com/api/v1/run/minimax/h3-max/reference-to-video?wait=20" \
  -H "Authorization: Bearer $DOITONG_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "prompt": "A red fox trots through fresh snow at sunrise, soft golden light, shallow depth of field",
    "prompt_expansion_mode": "balanced"
  }'

# Still "queued" or "running"? Poll with the id from the response:
curl "https://vit.doitong.com/api/v1/requests/REQUEST_ID?wait=20" \
  -H "Authorization: Bearer $DOITONG_API_KEY"

Phandel kava mangipen

Tiro kluč ačhel kadale browsereste: bičhalel pes numa ko Doitong API thaj šuvav pes (localStorage) numa te tiknes Remember. Le jekh API ključ

Palal H3 Max Reference to Video

Generinela video dirigime referencake asseturenca: dži ko 9 kipesa, 3 video thaj 3 audio klipura, buteder 12 fajlura savorenca. O prompto vakerel palal lende sar Image 1, Video 1, Audio 1 thaj aver. Opcionalne anglune, maškarutne (`middle_image_url` andro `middle_frame_time`) thaj palutne kadra pininena start, maškar thaj agor. `duration` si 0.92 dži ko 15 sekunda, thaj `aspect_ratio` šaj te avel adaptive

Parametra

Bičhalav vaj e kanonikano nav vaj e native nav, na duj. Kanonikane nava

NavTip standardnoMuklo Vakeripe
prompttrubul
Tipstring
≤ 50,000 simbolura
Video-generation instrukcije referencenca sikavde asset type taj position ande sako reference list, labelenca sar Audio 1, Video 1, Image 1, Image 2, taj avre.
reference_image_urlsopcionalno images
Tipurl[] · image
≤ 9 artiklura
Subject vaj style reference kipesko URL, sajekh 12-fajlo maximumesa savore video taj audio referencesa. Sikav e kipi ando prompto e listako positionesa: Image 1, Image 2, taj avre labels.
reference_video_urlsopcionalno
Tipurl[] · video
≤ 3 artiklura
URL adresi le miškipeske vaj aver referentne videonge, prinžarde palal i pozicija ande lista ando prompto pe etiketenca Video 1, Video 2 thaj aver pala lende. O khetano limito le kipenge, videonge thaj audio referencenge si 12 fajlura; sako video klipo trubul te avel 2–15 sekunde, thaj sa khetane na maj but sar 15 sekunde.
reference_audio_urlsopcionalno
Tipurl[] · audio
≤ 3 artiklura
De referentno audio sar URL adresi thaj prinžar le klipura ando prompto pe etiketenca Audio 1, Audio 2 thaj aver pala lende. Sako klipo trubul te avel 2–15 sekunde, thaj sasto audio na maj but sar 15 sekunde. Le kipenge, videonge thaj audio referencenge tipura šaj te den pes korkore vaj khetane, dži ko khetano limito 12 fajlura.
image_urlopcionalno image
Tipurl · image
Opcionalno kipesko URL savo sikavel o exacto anglutno kadro taj kerel o rezultatosko canvas. Reference assets arakhen e subjectosko consistency.
middle_image_urlopcionalno
Tipurl · image
Opcionalno guidance kipo vaš middle_frame_time, dostupno numa native 480P vaj 768P keripasa taj kana si delime angluno taj agorutno kipo.
middle_frame_timeopcionalno
Tipnumber | null
> 0< 15
Middle-image placement time, šaj te lelpe numa khetane middle_image_urlesa taj numa ande pozicije striktno maškar e mangline videoske anglune taj agorutne kadro. Del sekunda save nakhle katar o agor; o value si rounded ko maj pašesko 24 fps kadro.
end_image_urlopcionalno end_image
Tipurl · image
Opcionalno agorutno-kiposko URL savo fixinel o videosko exacto agorutno kadro pe delime kipos.
durationopcionalno
Tipnumber
standardno5
0.92 – 15
Manglo durata ande sekunda, savo mukel realno video dži 0.7 s maj dur sar manglo.
resolutionopcionalno
Tipenum
standardno768P
480P768P1080P
Keripasko rezolucija: 1080P mangela latentno dorikheripe e materialosko kerde native 768P-te, na native keripe ande avrirodli rezolucija.
seedopcionalno
Tipinteger | null
Bijočaripasko semeno vaš random keripe; te na delpes, jekh randomno avralas.
enable_safety_checkeropcionalno
Tipboolean
standardnotrue
Safety-checker aktivacia; true aktivarel checking.
prompt_expansion_modetrubul
Tipstring
standardnobalanced
Promptosko dorikheripasko zor angle keripe: 'quality' kerel jekh barvaleder prompto maj but sar 30s, taj 'balanced' delpes ande jekh sekunda. Avrirod 'disabled' te na kerel promptosko pherdipe.
aspect_ratioopcionalno
Tipenum
standardnoadaptive
adaptive21:916:94:31:13:49:16
Aspect ratio vaš o rezultatosko video.
reference_conditionsopcionalno
Tipobject[]
≤ 12 artiklura
Reference list delime sar alternativa vaš reference_*_urls, na khetane lenca. E ordering maškar modalities si významno, vaš kodo e sequence arakhelpe savore processingosko.
reference_conditions[].typetrubul
Tipenum
imagevideoaudio
Reference-ki modality; e ordering maškar modalities ande list si významno.
reference_conditions[].urltrubul
Tipstring
≥ 1 characters
Reference mediaso URL.
trubul Kanonikano nav, vi akceptime

Fajloske parametra lel jekh publikano https:// URL, jekh data: URI dži 10 MB, vaj e file_url uploadoski. Fajlura

Rezultato

GET /requests/req_8k2m4q7x1z9c3v5b6n0p2r4t
{
  "id": "req_8k2m4q7x1z9c3v5b6n0p2r4t",
  "object": "request",
  "model": "minimax/h3-max/reference-to-video",
  "status": "succeeded",
  "created_at": "2026-10-20T12:00:00Z",
  "started_at": "2026-10-20T12:00:02Z",
  "completed_at": "2026-10-20T12:00:09Z",
  "queue_position": null,
  "cost": {
    "credits": 400,
    "usd": "0.40",
    "charged": 400,
    "refunded": 0
  },
  "input": {
    "prompt": "A red fox trots through fresh snow at sunrise, soft golden light, shallow depth of field",
    "prompt_expansion_mode": "balanced"
  },
  "output": {
    "video": {
      "url": "https://content.doitong.com/dev-api/2026/10/20/req_8k2m4q7x1z9c3v5b6n0p2r4t/0-output.mp4",
      "content_type": "video/mp4",
      "file_name": "0-output.mp4",
      "file_size": 1843200
    },
    "expanded_prompt": "string",
    "seed": 42,
    "timings": {}
  },
  "outputs": [
    {
      "type": "video",
      "url": "https://content.doitong.com/dev-api/2026/10/20/req_8k2m4q7x1z9c3v5b6n0p2r4t/0-output.mp4",
      "content_type": "video/mp4",
      "file_size": 1843200,
      "width": 1280,
      "height": 720,
      "duration": 5,
      "stored": true
    }
  ],
  "error": null,
  "metrics": {
    "inference_seconds": 6.8
  },
  "urls": {
    "self": "https://vit.doitong.com/api/v1/requests/req_8k2m4q7x1z9c3v5b6n0p2r4t",
    "cancel": "https://vit.doitong.com/api/v1/requests/req_8k2m4q7x1z9c3v5b6n0p2r4t/cancel"
  }
}
Tire fajlura, tiro alosaripenSako outputs[].url si kopija amare arakhipneste: vično (čargeime per GB-mase) vaj 7 dives bilačhes, tiro kontoskere regulaciatar vaj ?storage=.
Sa jekh forma sako modeloθartype, url, content_type, width, height thaj duration, varesave imajinake, video thaj audio-ke. output si o modelosko korkoro resultosko objekto.
Outputoske fieldura
video
object {url, content_type, file_name, file_size}Video kerdo modelostar.
expanded_prompt
string | nullModeloske pherdo prompto, vaj null te naj pherdo teksto. E null situacije si disabled expansion, expansion savo na paruvdas, taj expansion savo o modelo hostoski API kerel andre.
seed
integerBase seed savo reprodukcir-el o kerdo rezultato.
timings
object | nullVaktiro meripe sikavdo ande sekunda, dostupno numa ande routes save sikaven backend timing; averčhane null. O 'inference' entry merel keci vakti lel e DiT denoising pe GPU backend.

Bute modeli: Referenca ko Video

AI agentenge: kadija rig sar Markdown si kate https://doitong.com/rom/developers/models/minimax/h3-max/reference-to-video.md. Sa e modelura thaj e gidura: https://doitong.com/developers/llms.txt