Saltar al cuntegna

VGGT-1B

Integrar

ID dal model vggt-1b
# Run the model; wait up to 20 s for the finished result in the same call
curl -X POST "https://api.doitong.com/v1/run/vggt-1b?wait=20" \
  -H "Authorization: Bearer $DOITONG_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{}'

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

Fin a $1,755 per execuziun pajà tenor l'utilisaziun effectiva, il rest vegn restituì Lantschar sin questa pagina

Obtegna ina clav API
Meta Vision

Reconstruescha ina scena 3D a partir da maximalmain 48 maletgs u in video e returna cartas da profunditad, posiziuns da camera e ina niv da puncts

Meta · Datas (JSON)

Tes emprim resultat vegn qua

Construis ina dumonda sutvart e l'exequescha cun tia clav API. Outputs vegnan copiads en nossa memorisaziun, conservads per adina u 7 dis gratuitamain.

Construir ina dumonda

Ils controls vegnan generads dal schema dal model. Il code e ​​il pretsch s'actualiseschan durant che ti als midas.

Avanzà images · video · alpha_blend_onto · array_encoding · frame_sampling_rate · confidence_percentile · max_points · export_prediction_data · export_depth_maps · export_point_cloud · enable_safety_checker · exclude_keys
image_urls
Trametta in datoteca u · fin a 48

Maletgs da funtauna per la reconstrucziun, mintgin cun padding en in quadrat e redimensiunà a ina dimensiun maximala da 518 pixels. JPEG, PNG, WEBP e mintga ulteriur format da maletg decodabel cun Pillow vegnan acceptads.

video_url
Trametta in datoteca u

Video da funta facultativ, dal qual ils frame campiunads suondan a image_urls en l'input cumponì. La campiunatura utilisescha mintga `frame_sampling_rate`-avel frame, includend adina l'emprim ed il davos frame.

alpha_blend_onto standard white

Il tractament dal chanél alpha porscha quatter opziuns: 'keep' allontanescha l'alpha senza midar ils valurs dals pixels originals; 'mean' cumpna cun la media RGB da ImageNet. Per in fund solid, tscherna 'white' u 'black' per cumpunar cun la colur correspundenta.

array_encoding standard base64

Format da memorisaziun dals arrays per las datotecas JSON da predicziun; 'list' dovra bler dapli spazi che 'base64'. Tscherna 'list' per glistas JSON annidadas u 'base64' per objects che cuntegnan {data, shape, dtype}.

standard 24

Interval dals frame applitgà mo sch'in video vegn furnì via `video_url`; autramain n'ha el nagin effect. Ils dus frame da cunfin vegnan mantegnids independentamain da l'interval, cun mintga n-avel frame campiunà tranteren.

standard 50

Percentuala da filtraziun per la construcziun dal point cloud basada sin la confidenza: 0 deactivescha la filtraziun, auters valurs excludan quella percentuala da points, cumenzond cun ils main confidents.

standard 250000

Limit dal dumber da points GLB, standardmain 250,000. Per reducir la latenza da serialisaziun e da l'upload, vegn scritta ina selecziun campiunada regularmain, sch'il cloud surpassa quest budget.

Clavs exclusas da mintga datoteca JSON da predicziun.

POST /run/vggt-1b?wait=20 calculaziun → $1,755
# Run the model; wait up to 20 s for the finished result in the same call
curl -X POST "https://api.doitong.com/v1/run/vggt-1b?wait=20" \
  -H "Authorization: Bearer $DOITONG_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{}'

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

Executar questa dumonda

Tia clav resta en quest browser: ella vegn tramessa mo al Doitong API e memorisada (localStorage) mo sch'ella scheccas Tiar a memoria. Obtegnair ina clav API

Davart VGGT-1B

Stimescha la geometria a partir da pliras perspectivas. Dà maximalmain 48 maletgs en `image_urls` e, opziunalmain, agiunta in `video_url` campionà mintga `frame_sampling_rate` frames. Retorna cartas da profunditad per mintga frame, extrinsics e intrinsics da la camera, datotecas da predicziun bruttas e ina niv da puncts GLB cun conus da camera. `confidence_percentile` e `max_points` controllan la densitad da la niv da puncts.

Parameters

Invia u il num canonic u il num nativ, betg omaduas. Nums canonics

NumTip StandardPermess Descripziun
image_urlsopziunal images
Tipurl[] · image
≤ 48 elements
Maletgs da funtauna per la reconstrucziun, mintgin cun padding en in quadrat e redimensiunà a ina dimensiun maximala da 518 pixels. JPEG, PNG, WEBP e mintga ulteriur format da maletg decodabel cun Pillow vegnan acceptads.
video_urlopziunal video
Tipurl · video
Video da funta facultativ, dal qual ils frame campiunads suondan a image_urls en l'input cumponì. La campiunatura utilisescha mintga `frame_sampling_rate`-avel frame, includend adina l'emprim ed il davos frame.
frame_sampling_rateopziunal
Tipinteger
Standard24
1 – 1 000
Interval dals frame applitgà mo sch'in video vegn furnì via `video_url`; autramain n'ha el nagin effect. Ils dus frame da cunfin vegnan mantegnids independentamain da l'interval, cun mintga n-avel frame campiunà tranteren.
alpha_blend_ontoopziunal
Tipenum
Standardwhite
keepwhiteblackmean
Il tractament dal chanél alpha porscha quatter opziuns: 'keep' allontanescha l'alpha senza midar ils valurs dals pixels originals; 'mean' cumpna cun la media RGB da ImageNet. Per in fund solid, tscherna 'white' u 'black' per cumpunar cun la colur correspundenta.
export_prediction_dataopziunal
Tipboolean
Standardtrue
Exportar las predicziuns bruttas sco datotecas JSON separadas per frame singuls. Mintga datoteca cuntegna la profunditad e sia confidenza, ina mascra e la posiziun da camera codada.
export_depth_mapsopziunal
Tipboolean
Standardtrue
Exportar las raster da profunditad cun ina mappatura lineara da `depth_range` als valurs da pixels 0–65535. Mintga frame survegn ina PNG monocroma separada cun precis 16 bits.
export_point_cloudopziunal
Tipboolean
Standardtrue
Exportar il point cloud colurà fusiunà en il format GLB, cun las cameras stimadas represchentadas sco meshes conicas singulas en la scena.
array_encodingopziunal
Tipenum
Standardbase64
base64list
Format da memorisaziun dals arrays per las datotecas JSON da predicziun; 'list' dovra bler dapli spazi che 'base64'. Tscherna 'list' per glistas JSON annidadas u 'base64' per objects che cuntegnan {data, shape, dtype}.
exclude_keysopziunal
Tipenum[]
Clavs exclusas da mintga datoteca JSON da predicziun.
confidence_percentileopziunal
Tipnumber
Standard50
0 – 100
Percentuala da filtraziun per la construcziun dal point cloud basada sin la confidenza: 0 deactivescha la filtraziun, auters valurs excludan quella percentuala da points, cumenzond cun ils main confidents.
max_pointsopziunal
Tipinteger
Standard250000
1 000 – 5 000 000
Limit dal dumber da points GLB, standardmain 250,000. Per reducir la latenza da serialisaziun e da l'upload, vegn scritta ina selecziun campiunada regularmain, sch'il cloud surpassa quest budget.
enable_safety_checkeropziunal
Tipboolean
Standardtrue
Activaziun dals controlls da segirezza per maletgs ed videos d'input.
obligatori Num canonic, er acceptà

Ils parameters da datotecas acceptan ina URL publica https://, ina URI data: fin a 10 MB u il file_url d'ina chargiada. Datotecas

Output

GET /requests/req_8k2m4q7x1z9c3v5b6n0p2r4t
{
  "id": "req_8k2m4q7x1z9c3v5b6n0p2r4t",
  "object": "request",
  "model": "vggt-1b",
  "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": 1755,
    "usd": "1.755",
    "charged": 1755,
    "refunded": 0
  },
  "input": {},
  "output": {
    "point_cloud": {
      "url": "https://content.doitong.com/dev-api/2026/10/20/req_8k2m4q7x1z9c3v5b6n0p2r4t/0-output.jpeg",
      "content_type": "image/jpeg",
      "file_name": "0-output.jpeg",
      "file_size": 1843200
    },
    "depth_maps": [
      {
        "url": "https://content.doitong.com/dev-api/2026/10/20/req_8k2m4q7x1z9c3v5b6n0p2r4t/0-output.jpeg",
        "content_type": "image/jpeg",
        "file_name": "0-output.jpeg",
        "file_size": 1843200,
        "width": 1024,
        "height": 768
      }
    ],
    "prediction_data": [
      {
        "url": "https://content.doitong.com/dev-api/2026/10/20/req_8k2m4q7x1z9c3v5b6n0p2r4t/0-output.jpeg",
        "content_type": "image/jpeg",
        "file_name": "0-output.jpeg",
        "file_size": 1843200
      }
    ],
    "num_frames": 1,
    "depth_range": [
      1.0
    ],
    "extrinsics": [
      [
        [
          1.0
        ]
      ]
    ],
    "intrinsics": [
      [
        [
          1.0
        ]
      ]
    ],
    "timings": {}
  },
  "outputs": [
    {
      "type": "file",
      "url": "https://content.doitong.com/dev-api/2026/10/20/req_8k2m4q7x1z9c3v5b6n0p2r4t/0-output.jpeg",
      "content_type": "image/jpeg",
      "file_size": 1843200,
      "stored": true
    }
  ],
  "error": null,
  "metrics": {
    "inference_seconds": 6.8
  },
  "urls": {
    "self": "https://api.doitong.com/v1/requests/req_8k2m4q7x1z9c3v5b6n0p2r4t",
    "cancel": "https://api.doitong.com/v1/requests/req_8k2m4q7x1z9c3v5b6n0p2r4t/cancel"
  }
}
Tias datotecas, tia decisiunMintga outputs[].url è ina copia en nossa memorisaziun: conservada per adina (pajà per GB-mes) u 7 dis gratuitamain, tenor la configuraziun dal conto u ?storage=.
La medema structura per mintga modeltype, url, content_type, width, height e duration, tant per maletgs sco per video e audio. output è l'object da resultat dal model sez.
Champs da l'output
point_cloud
object {url, content_type, file_name, file_size} | nullExport dal point cloud restituì, sch `export_point_cloud` è true; sia scena GLB cumbinescha il cloud fusiunà cun conus da camera.
depth_maps
array of object {url, content_type, file_name, file_size, width, height}Export da la carta da profunditad disponibel, sch `export_depth_maps` è true. Las entradas suondan la sequenza dals frame da funtauna, cun mintga frame represchentà sco raster PNG monocroma cun profunditad da 16 bits.
prediction_data
array of object {url, content_type, file_name, file_size}Export da las predicziuns bruttas: sch `export_prediction_data` è true, ha mintga frame sia atgna datoteca JSON, e las datotecas suondan la sequenza dals frame.
num_frames
integerDumber total da frame reconstruids, inclus maletgs e frame campiunads dal video.
depth_range
array of numberLimites da profunditad [min, max] utilisads per la quantisaziun da `depth_maps`. Recuverta la profunditad sco min + pixel * (max - min) / 65535.
extrinsics
array of array of array of numberMatrices extrinsicas camera-da-mund, en la grondezza 3x4, per mintga frame; ils ass OpenCV mussan a dretga per x, engiu per y e davant per z.
intrinsics
array of array of array of numberMatrices intrinsicas pinhole, en la grondezza 3x3, per mintga frame, exprimidas en coordinatas da pixels dal frame dal model da 518x518.
timings
objectTemp passà per mintga stadi, en secundas.

Per agents AI: questa pagina sco Markdown è disponibla sut https://doitong.com/rm/developers/models/vggt-1b.md. Tut ils models e las guidas: https://doitong.com/developers/llms.txt