3.1.1   static

deprecated

Static bald head with separate haircuts.

3.1.2   animated_face

Bald head with the optional set of blendshapes for animations (45 facial, 17 visemes) and separate haircuts.

Available subtypes:

• base/legacy - static bald head with a set of blendshapes (45 facial, 17 visemes) for animations and separate haircuts. Default if no subtype specified
• indie/legacy_styled subtype have different model topology to make available optional mesh modifications (see computation parameters section below)

Model sample

Available LODs:

Note: The structure of the head in LOD8 is the same as in LOD7, but without teeth.

3.1.3   head_1.0

deprecated

Predicted head, haircut, and bust. All-in-one model with unique topology.

3.1.4   head_1.1

deprecated

Predicted head, haircut, and bust. All-in-one model with unique topology. Improved texture and mesh quality (compared to head_1.0).

3.1.5   head_1.2

Predicted head, haircut, and bust. All-in-one model with unique topology. Fixed shape and texture artifacts in long hair and busts that are present in head_1.1. Supports export parameters (multiple formats, unified or separate meshes, textures, etc).

Available subtypes:

• base/static - static predicted head, haircut, and bust
• base/legacy - predicted head, haircut, and bust with a set of blendshapes (45 facial, 17 visemes) for animations
• base/mobile - predicted head, haircut, and bust with a set of blendshapes (51 facial) for animations

Model samples

Available LODs:

3.1.6   head_2.0

Takes the best of head_1.1 and animated_face avatars. Looks exactly like head_1.1, but a hairstyle is represented with a separate mesh (except for bust/static subtype), and a bust can be removed (head/mobile subtype). Supports export parameters (multiple formats, unified or separate meshes, textures, etc).

Please note: available on the Pro subscription plan only.

Available subtypes:

• bust/mobile - predicted bust with a detachable haircut and set of 51 facial blendshapes
• bust/static - static predicted bust and haircut
• head/mobile - predicted head with a detachable haircut and set of 51 facial blendshapes. The mesh topology of a head is always the same

Model samples

Available LODs:

3.1.7   body_0.1

deprecated

Predicted bald head and full body model with fixed topology and detachable predicted haircut.

3.1.8   body_0.2

Predicted bald head and full body model with fixed topology, predicted and artificial haircuts, and artificial outfits. Predictions of the body shape may be improved by specifying the height, weight, gender, and body measurements of the model. Model skeleton is also available for use with animations.

Please note: the pipeline supports adult bodies only. If you specify parameters as for a child (small height and small weight) then the pipeline will return a body following these parameters, but it will be the body of a small adult, not a child.

Please note: available on the Pro subscription plan only.

Available subtypes:

• body/mobile - full body model with a set of 51 facial blendshapes and 15 visemes.

Model samples

3.1.9   body_0.3

Evolution of body_0.2 pipeline: improved prediction of the body shape and support of various export options via the export parameters (multiple formats, unified or separate meshes, textures, etc).

Please note: available on the Pro subscription plan only.

Please note: this pipeline utilizes export parameters

Available subtypes:

• mobile (FitPerson) - full body model based on the SMPL model that is created by the Max-Planck Institute with a set of 51 facial blendshapes and 15 visemes. Have an additional head export template, which is effectively head/mobile subtype of head_2.0 pipeline.
• female, male (MetaPerson) - a set of bodies created by 3D artists with a head automatically generated from selfies. The animation rig for the bodies is compatible with Adobe's Mixamo, which makes them compatible with a large existing set of animations. Fixed body shape allows 3D artists to apply custom designed outfits easier than before. The head is based on the head_2.0 pipeline.

Model samples

Available LODs:

• FitPerson

  	default template		head template
  LOD#	vertices	faces	vertices	faces
  0	18689	34578	12367	23404
  1	15633	28102	10374	19054
  2	13850	24880	9116	16860
  3	9285	16173	6059	11007
  4	8096	13905	5287	9497
  5	6987	11565	4587	7889
  6	6271	10151	4106	6937
  7	4267	7140	2102	3926

• MetaPerson

  	female/male subtype
  LOD#	vertices	faces
  0	19538	36196
  1	15633	28102
  2	13854	24880
  3	9291	16173
  4	8102	13905
  5	6993	11565
  6	6277	10151
  7	4273	7140

3.3.1   Haircuts

We have three sets of artificial haircuts which are modified during avatar computation process to match each particular avatar head geometry:

base set

3 male and 3 female haircuts

Haircuts samples

facegen set

39 haircuts in total, some of them are unisex

Haircuts samples

plus set

50 haircuts in total, some of them are unisex

Haircuts samples

Availability of these sets depends on your current pricing plan.

Also we have additionally generated haircut. This haircut is unique and generated for each particular avatar to match the haircut from the source photo. Please note: if there is detected that the source photo does not have a haircut (i.e. a bald person) - there will be no generated haircut in the list of available haircuts for the computed avatar.

It is also possible to add your haircut: contact us at support@avatarsdk.com

3.3.2   Outfits

Avatar SDK provides artificial outfits for full body avatars. Each outfit consists of mesh, texture, body visibility mask, roughness, metalness, and normals map. Alpha channel from the body visibility mask should be applied to the avatar body such that the avatar body becomes transparent to avoid body and outfit overlap during avatar animation. Please note: each outfit is provided with a separate body visibility mask, thus the corresponding mask should be applied.

base set

Set of business, business casual, and casual outfits

Outfits samples

lowpoly set

The same outfits set as the base but with lowered polygons count. Outfit name ends with _lowpoly suffix. Please see the outfit mesh polygons comparison table below for more details about 3D mesh.

meta set

The same outfits set as the base but adapted and available only for MetaPerson avatars. Outfit name contains _meta_ suffix (e.g. outfit_meta_0). Please see the outfit mesh polygons comparison table below for more details about 3D mesh.

Available outfits 3D mesh details:

Please see parameters retrieval request (or enumerable export parameters retrieval request for body_0.3 pipeline) to retrieve a list of available outfits.

3.3.3   Blendshapes

Avatar SDK supports animated avatars using blendshapes. There are three different blendshapes sets available which should be used depending on your use-case: general animation or lip-sync. Please consult the blendshapes set description for more information. Please note that the legacy blendshapes set is not recommended to use in the new application as it will be deprecated eventually.

legacy_45

The basic set of 45 facial blendshapes

mobile_51

Set of 51 facial blendshapes. This set is compatible with Apple ARKit: https://developer.apple.com/documentation/arkit/creating_face-based_ar_experiences

visemes_17

Set of 15 visemes (AA, CH, EE, IH, OH, OU, TH, dd, ff, kk, nn, pp, rr, sil) and two additional blendshapes: frown and smile. This set is compatible with Oculus Lipsync that required these 15 visemes. You can find more details here: https://developer.oculus.com/documentation/audiosdk/latest/concepts/book-audio-ovrlipsync/

Please find our guides on how to use visemes for lipsync here: https://docs.avatarsdk.com/unity-plugin/2.0.0/lipsync_samples.html

visemes_15

Set of 14 visemes (PP, FF, TH, DD, kk, CH, SS, nn, RR, aa, E, ih, oh, ou) and one additional blendshape: Laughter. This set is identical to Oculus Lipsync visemes. You can find more details here: https://developer.oculus.com/documentation/audiosdk/latest/concepts/book-audio-ovrlipsync/

3.3.4   Model info

Avatar SDK is capable to provide some meta-information about computed avatar from source photo like haircut, skin, eyes color, etc. Please find the list of available meta-information parameters below:

age

Classify a person's age from a submitted photo. Currently possible values are child and not_child. age_confidence also will be present in model_info to represent the confidence of age classification with value in the range [0.5, 1] inclusive

eye_iris_color

Compute average eye iris color [1] from a submitted photo

eye_sclera_color

Compute average eye sclera color [1] from a submitted photo

facial_landmarks_68

Compute facial landmarks. Landmark coordinates are represented as flat JSON array with X, Y, Z float components sequence for each landmark, or NaN (for each component) if no landmark detected

Please note: for base/legacy and indie/legacy_styled subtypes of animated_face pipeline, landmarks in range [0-7] and [9-16] are not present (i.e. they set to NaN).

Landmarks correspondence

(click on image to magnify)

gender

Predict a person's gender from a submitted photo. Possible values are male and female. gender_confidence also will be present in model_info to represent the confidence of gender prediction with value in the range [0.5, 1] inclusive

hair_color

Compute average haircut color [1] from a submitted photo

lips_color

Compute average lips color [1] from a submitted photo

predict_haircut

Predict which haircut from the base and the facegen sets (base and plus sets for head_2.0) matches best to the submitted photo. haircut_name will contain the corresponding haircut identifier. May contain special value absolutely_bald which represents the absence of haircut. Please note: predicted haircut is independent on which haircuts were requested to compute with a particular avatar

For head_2.0 additionally estimates haircuts correspondence and put weights under the haircut_full_info key. Please see the sample model info below:

{
   "haircut_full_info" : {
      "absolutely_bald" : 2.936517375928815e-06,
      "afro_short" : 0.0112605718895793,
      "balding" : 0.002860783133655787,
      "bob" : 0.02034652791917324,
      "bob_parted" : 0.03573336824774742,
      "burr_cut" : 0.009195085614919662,
      "classic_side_part" : 0.0149933909997344,
      "classic_tapper" : 0.009302150458097458,
      "cowlick" : 0.005376111250370741,
      "cowlick2" : 0.012983413413167,
      "crew_cut" : 0.01081948075443506,
      "crew_cut2" : 0.01268314570188522,
      "crew_cut3" : 0.01463177613914013,
      "curtained_hair" : 0.03242894262075424,
      "flattop" : 0.01193561963737011,
      "high_volume_brushed_up" : 0.01360709313303232,
      "long_afro" : 0.02262603491544724,
      "long_bob" : 0.04952326044440269,
      "long_crimped" : 0.05819306150078773,
      "long_disheveled" : 0.06030699983239174,
      "long_hair" : 0.06073638796806335,
      "long_hair2" : 0.075968898832798,
      "long_wavy" : 0.06359551101922989,
      "mid_length_ruffled" : 0.02399636991322041,
      "mid_length_straight" : 0.04574619233608246,
      "mid_length_wispy" : 0.03018877282738686,
      "ponytail_with_bangs" : 0.03226388990879059,
      "short_bob_asymmetrical_bangs" : 0.02550809271633625,
      "short_curls2" : 0.02223772928118706,
      "short_disheveled" : 0.01950234733521938,
      "short_simple" : 0.01544607523828745,
      "side_french_braid" : 0.04054473340511322,
      "straight_bob_bangs" : 0.03596006706357002,
      "very_long" : 0.06915491819381714,
      "wavy_shag" : 0.02055038698017597
   },
   "haircut_name" : "long_hair2",
   "pipeline" : "head_2.0",
   "pipeline_subtype" : "bust/mobile",
   "race" : "white",
   "race_confidence" : 0.9999996423721313,
   "races" : {
      "asian" : 2.615514413124159e-13,
      "black" : 4.136433915391535e-07,
      "white" : 0.9999996423721313
   }
}

race

Predict person race from the submitted photo. Currently, three races are predicted (asian, black and white) with certain confidence for each race in the range [0, 1] inclusive. The resulting information will include the most probable race with its confidence and a JSON Object with all races with its confidences. Please see the sample model info below:

{
   "pipeline" : "head_2.0",
   "pipeline_subtype" : "bust/mobile",
   "race" : "white",
   "race_confidence" : 0.9999996423721313,
   "races" : {
      "asian" : 2.615514413124159e-13,
      "black" : 4.136433915391535e-07,
      "white" : 0.9999996423721313
   }
}

skin_color

Compute average skin color [1] from a submitted photo

Colors correspondence

3.3.5   Avatar modifications

Avatar SDK is aimed to produce recognizable photo-realistic avatars. However, there is demand on avatars modifications, leaving avatar recognizable, though. Please find available avatar modifications below.

Please note: all parameters for this category are key-value, and modifies the original avatar respectively its description.

add_eyelid_shadow : boolean

Add an eyelid shadow on top of eye texture. Works only when parametric_eyes_texture is set. See textures comparison below

Models comparison

add_glare : boolean

Add a glare directly into parametric eye texture. Works only when parametric_eyes_texture is set. Enabled by default. See textures comparison below

Models comparison

allow_modify_neck : boolean

By default, all avatars have the same neck mesh across all models. If flag set to true, neck mesh will be modified to better match the submitted photo. Please note: x-axis rotation is also applied to match a head pose from submitted photo

Models comparison

caricature_amount : float

Factor to strengthen differences between average model and a particular avatar. Valid values range is [0, +∞). Usable values range is model dependent and usually should be lower than 5

Models comparison

curved_bottom : boolean

Make the bottom of the model bust slightly curved. See models comparison below

Models comparison

enhance_lighting : boolean

Attempt to make lighting more uniform. See models comparison below

Models comparison

eye_iris_color : color [1]

Recolor eye iris directly on a model texture. Works only when parametric_eyes_texture is set. See textures comparison below for {"red": 0, "green": 255, "blue": 0} color

Models comparison

eye_sclera_color : color [1]

Recolor eye sclera directly on a model texture. Works only when parametric_eyes_texture is set. See textures comparison below for {"red": 0, "green": 255, "blue": 0} color

Models comparison

gender : deprecated

Please note: this parameter is deprecated, please use the gender parameter in the body shape parameters section for body_0.2 pipeline.

Expected model gender. Available values are male and female. If not specified, gender is autodetected.

Models comparison

generated_haircut_faces_count : integer

Resample generated haircut mesh to have specified faces count. See models comparison below for default and 256 faces count haircut (rendered without texture)

Models comparison

generated_haircut_texture_size : size [2]

Resize generated haircut texture to the specified size. Original haircut texture size is 1024x1024 px, thus setting size greater than this value will not result in improved texture quality.

hair_color : color [1]

Recolor generated haircut directly on a haircut texture. See models comparison below for {"red": 0, "green": 255, "blue": 0} color

Models comparison

height : deprecated

Please note: this parameter is deprecated, please use the height parameter in the body shape parameters section for body_0.2 pipeline.

Expected model height in meters, greater than zero. Please note: does not take effect without weight specified.

Models comparison

lips_color : color [1]

Recolor lips directly on a model texture. See models comparison below for {"red": 0, "green": 255, "blue": 0} color

Models comparison

parametric_eyes_texture : boolean

Replace eye texture from the submitted photo with generated one with sclera and iris colors to match the photo version. Please note: always set for the indie/legacy_styled subtype of the animated_face pipeline. See models comparison below

Models comparison

parametric_eyes_texture_v2 : boolean

Replace eye texture from the submitted photo with generated one with sclera and iris colors to match the photo version. See models comparison below.

Please note: add_glare and add_eyelid_shadow flags are ignored when this flag is set to true.

Please note: If both parametric_eyes_texture and parametric_eyes_texture_v2 are set to true, then parametric_eyes_texture_v2 is ignored.

Models comparison

remove_smile : boolean

Attempt to remove a smile on the submitted photo before the actual avatar computation start. Currently detects and removes mostly open smiles (e.g. smiles with mouth open). The status of a smile removal may be obtained later from model information JSON Object by the remove_smile key (see avatar model information retrieval request). Please find the model comparison below.

Please note: modifies original photo, thus avatars computed on the same source photo with and without this flag may appear slightly different.

Models comparison

remove_stubble : boolean

Attempt to remove a stubble on the submitted photo before the actual avatar computation start. The status of a stubble removal may be obtained later from model information JSON Object by the remove_stubble key (see avatar model information retrieval request). Please find the model comparison below.

Please note: modifies the original photo, thus avatars computed on the same source photo with and without this flag may appear slightly different.

Models comparison

remove_glasses : boolean

Attempt to remove glasses on the submitted photo before the actual avatar computation start. Status of glasses removal may be obtained later from model information JSON Object by remove_glasses key (see avatar model information retrieval request). Please find the model comparison below

Models comparison

repack_texture : boolean

Optimize texture layout to occupy less space. See textures comparison below

Textures comparison

slightly_cartoonish_texture : boolean

Make the model texture look slightly cartoonish. Please note: replaces original avatar texture. See models comparison below

Models comparison

teeth_color : color [1]

Recolor teeth directly on a model texture. See models comparison below for {"red": 0, "green": 255, "blue": 0} color

Models comparison

texture_size : size [2]

Resize model texture to the specified size. Original texture size is 2048x2048 px, thus setting size greater than this value will not result in improved texture quality.

weight : deprecated

Please note: this parameter is deprecated, please use the weight parameter in the body shape parameters section for body_0.2 pipeline.

Expected model weight in kilograms, greater than zero. Please note: does not take effect without height specified.

Models comparison

3.3.6   Shape modifications

Similar to avatar modification parameters, see description for that parameters category.

Please note: all parameters for this category are key-value, and modify the original avatar respectively its description.

cartoonish_v0.3 : float

Factor to modify avatar mesh to look more cartoon-like. Valid values range is [0, +∞). Usable values range is model dependent and usually should be lower than 1

Models comparison

cartoonish_v1.0 : float

Factor to modify avatar mesh to look more cartoon-like. Valid values range is [0, +∞). Usable values range is model dependent and usually should be lower than 1

Models comparison

3.3.7   Body shape

Avatar SDK algorithms predict full body shape by analyzing an input photo and producing an average body for detected parameters. You can optionally specify additional parameters, such as gender, weight, height, etc. Parameters that were not specified will be predicted from an input photo. However, it is recommended to specify as many parameters as possible in order to get a specific body shape.

Please note: the output model may not necessarily have exactly the same measurements as provided by the avatar request. The output measurements are expected to be close to the input body shape parameters that are within the intervals specified by the tables below. Values outside of these intervals may result in significantly different output measurements. We are working to expand these intervals, specifically the one corresponding to the height parameter.

For body_0.2 pipeline:

Parameter name	Gender
	female	male	non_binary
height	[150; 171]	[160; 187]	[150; 187]
weight	[53; 76]	[68; 90]	[53; 90]
chest	[90; 116]	[96; 123]	[90; 123]
waist	[73; 99]	[80; 115]	[73; 115]
hips	[92; 119]	[96; 118]	[92; 119]

For body_0.3 pipeline (FitPerson only):

Parameter name	Gender
	female	male	non_binary
height	[150; 174]	[160; 187]	[150; 187]
weight	[54; 86]	[70; 112]	[54; 112]
chest	[90; 115]	[96; 125]	[90; 125]
waist	[73; 98]	[81; 118]	[73; 118]
hips	[90; 116]	[93; 119]	[90; 119]

Please note: all parameters for this category are key-value, and modify the original avatar respectively to their description.

gender : string

Expected model's gender. Available values are female, male, and non_binary. If not specified, gender is autodetected. It is highly recommended to specify this parameter to avoid incorrectly autodetected value.

height : float

Expected model's height in centimeters, greater than zero. If not specified, height is predicted. It is highly recommended to specify this parameter to avoid incorrectly autodetected value.

weight : float

Expected model's weight in kilograms, greater than zero. If not specified, weight is predicted. It is highly recommended to specify this parameter to avoid incorrectly autodetected value.

chest : float

Expected model's chest girth in centimeters, greater than zero. If not specified, chest girth is predicted.

waist : float

Expected model's waist girth in centimeters, greater than zero. If not specified, waist girth is predicted.

hips : float

Expected model's hips girth in centimeters, greater than zero. If not specified, hips girth is predicted.

3.3.8   Additional textures

You could request to compute additional textures along with default one, which is always computed. See avatar textures to find out how to retrieve additional textures.

cartoonish_texture

Uniformly colored texture

Textures comparison

lips_mask

Mask in PNG format

Textures comparison

metallic_map

Metalness map in JPG format

Textures comparison

roughness_map

Roughness map in JPG format

Textures comparison

slightly_cartoonish_texture

Slightly smoother than the original texture

Textures comparison

smooth_eyelashes_texture

Original texture with a smoothed area of eyelashes

Textures comparison

3.4.1   Available flags

additional_textures : list

List of textures to export besides default diffuse model texture. Available values are normal_map, roughness_map, metallic_map.

For outfits section body_visibility_mask is also available (see outfits section for visibility mask information).

For head_1.2 and head_2.0 pipelines lips_mask is also available for top-level additional_textures

color : color [1]

If texture support recoloring, this flag specifies the base color for the recoloring process

embed : boolean

If avatar format (i.e. top-level format) supports sub-meshes (i.e. fbx, gltf, glb) this flag indicates whether to embed this particular category meshes into the avatar or store these meshes as separate files. If avatar format does not support sub-meshes - flag is set to false and category meshes are saved as separate files

embed_textures : boolean

If format supports textures embedding (i.e. fbx or glb), this flag indicates whether to embed textures or store them as separate files. If avatar format does not supports textures embedding - flag is set to false and textures are saved as separate files

format : string

File format for the resulting mesh. The top-level flag defines the format of the avatar, category-level flag defines the format of the mesh for this particular category. Currently supported options include gltf, glb, fbx, obj, ply, and bin (blendshapes only)

list : list

List of enumerable values of assets to export for this particular category. Please see the enumerable export parameters retrieval request to retrieve a list of available enumerable export parameters for each particular category for pipeline and pipeline subtype on your account

pointclouds : boolean

If set to true, then the resulting model file will not contain mesh faces and UV-mapping (i.e. it will contain only mesh vertices).

For meshes with regular topology (i.e. models where a number of vertices, edges, and how these vertices connected are the same, only vertices positions are different), it may be convenient to download full mesh only once and then download and replace only pointclouds (only vertices) for all future models to reduce the download data size

texture_size : size [2]

Resize texture to the specified size. Avatar texture size is 4096x2048 px. Most asset textures are 4096x4096 px for outfits and 2048x2048 px for haircuts, but size varies between assets. Please check the original texture size before setting a custom value as setting it greater than the original one will not result in improved texture quality.

Please note: if texture size is also set in avatar modification parameters - then texture resized according to avatar modifications first, and then according export parameters.

template : string

Apply particular avatar export template. Currently, only one additional template is supported: head for body_0.3 pipeline, which is effectively head/mobile subtype of head_2.0 pipeline.

lod : integer

Level of details for exported mesh. Please find more information on mesh properties in the LOD table for each particular pipeline description in available pipelines section.

3.4.2   Export files

When export parameters are specified, algorithms generate export which consists of one or several export files depending on specified export parameters flags. In order to download all avatar's data, all files within one export have to be downloaded. Export file names are unique within one export and consistent from avatar to avatar export thus you may set priority to download export files in your application: e.g. avatar itself first, then its haircuts, etc.

Please see the avatar exports list request to find out how to list available avatar exports and their export files.

3.4.3   Static export files

Some export files may contain files, which are not modified from export to export: e.g. normals map or default, non-recolored haircut textures. These files are explicitly listed within each export file. They may be downloaded only once and cached on the client-side for other export files to save bandwidth and minimize export download size. Please note: they still have to be downloaded separately from the export file for the first time as they are not included in the export file archive.

Please see avatar exports list request description to find information about static files' retrieval.

4.1.1   List

Method

GET

URL

/players/

Description

Retrieve a paginated list of registered players available for your developer account. Available only in developer access mode. See the pagination section for paginated response details. See response example below:

[
  {
    "url":"https://api.avatarsdk.com/players/9418fd52-baef-401c-a2b3-e6fe16d93596/",
    "code":"9418fd52-baef-401c-a2b3-e6fe16d93596",
    "created_on":"2017-04-03T04:08:15.589833Z",
    "comment":""
  },
  ...
]

Filters

• created_before - filter Players with creation datetime before specified one (inclusive)
• created_after - filter Players with creation datetime after specified one (inclusive)
• comment - filter by part of the Player comment

4.1.2   Retrieve

Method

GET

URL

/players/{code}/

Description

Retrieve a particular instance of Player identified by code URL part. See player DTO for the response format description. Available in developer and client access modes.

4.1.3   Create

Method

POST

URL

/players/

Description

Register a new Player. See player DTO for response format description. Available in developer and client access modes.

Request parameters

No request parameters are required.

4.2.1   Retrieve

Method

GET

URL

/parameters/(available | default)/{pipeline}/

Description

Retrieve all available or default parameters for specified pipeline and default pipeline subtype. Use filters to retrieve parameters for the non-default pipeline subtype. The response is one key-value pair with pipeline_subtype as a key and parameters as a value.

Parameters may be just copy-pasted as-is (or partially) into parameters field of create avatar request parameters. The only exception is the avatar_modifications parameters category: this category contains key-value parameters. See the avatar modification parameters section for details.

Please see example response below (a little bit shortened for the sake of simplicity):

{
  "base/legacy": {
    "blendshapes": {
      "base": [
        "legacy_45",
        "visemes_17"
      ]
    },
    "haircuts": {
      "base": [
        "female_NewSea_J086f",
        "female_NewSea_J096f",
        "female_NewSea_J123f",
        "male_NewSea_J003m",
        "male_NewSea_J082m",
        "male_makehuman_short02"
      ],
      "facegen": [
        "facegen_afro",
        "facegen_balding",
        ...
        "facegen_wavy_shag"
      ]
    }
  }
}

Filters

• pipeline_subtype - filter by non-default pipeline subtype. See available pipelines for details about available pipeline subtypes

4.3.1   Retrieve

Method

GET

URL

/export_parameters/available/{pipeline}/

Retrieve all available enumerable export parameters for specified pipeline and pipeline subtype. Use filters to retrieve parameters for the non-default pipeline subtype. The response is JSON object of available enumerable values for each export parameters category.

Please see example response below (a little bit shortened for the sake of simplicity):

{
    "blendshapes": [
        "mobile_51",
        "visemes_15"
    ],
    "haircuts": [
        "generated",
        ...
        "wavy_bob"
    ]
}

Filters

• pipeline_subtype - filter by subtype. See available pipelines for details about available pipeline subtypes

4.4.1   List

Method

GET

URL

/avatars/

Description

Retrieve a paginated list of available avatar models for your account. Available in developer and client access modes. In the client access mode list of avatars is restricted to models created by specified PlayerUID. See the pagination section for paginated response details. See response example below:

[
  {
    "url":"https://api.avatarsdk.com/avatars/b63f9737-8594-460a-99a4-74e60b042bd4/",
    "code":"b63f9737-8594-460a-99a4-74e60b042bd4",
    "status":"Completed",
    "progress":100,
    "name":"Leo",
    "pipeline": "animated_face",
    "description":"",
    "created_on":"2017-04-03T04:08:15.589833Z",
    "ctime":null,
    "mesh":"https://api.avatarsdk.com/avatars/b63f9737-8594-460a-99a4-74e60b042bd4/mesh/",
    "texture":"https://api.avatarsdk.com/avatars/b63f9737-8594-460a-99a4-74e60b042bd4/texture/",
    "thumbnail":"https://api.avatarsdk.com/avatars/b63f9737-8594-460a-99a4-74e60b042bd4/thumbnail/",
    "haircuts":"https://api.avatarsdk.com/avatars/b63f9737-8594-460a-99a4-74e60b042bd4/haircuts/",
    "blendshapes":"https://api.avatarsdk.com/avatars/b63f9737-8594-460a-99a4-74e60b042bd4/blendshapes/"
  },
  ...
]

Filters

• name - filter by part of the avatar name
• description - filter by part of the avatar description
• status - filter by part of the avatar status
• created_before - filter avatars with creation datetime before specified one (inclusive)
• created_after - filter avatars with creation datetime after specified one (inclusive)
• changed_before - filter avatars with change datetime before specified one (inclusive)
• changed_after - filter avatars with change datetime after specified one (inclusive)
• pipeline - filter by used pipeline

4.4.2   Retrieve

Method

GET

URL

/avatars/{code}/

Description

Retrieve particular instance of avatar identified by code URL part. See avatar DTO for response format description. Available in developer and client access modes.

4.4.3   Create

Method

POST

URL

/avatars/

Description

Create a new avatar calculation task. Available in developer and client access modes. After POSTing a request, avatar photo for the calculation is being uploaded to calculation backend (state Uploading), then Queued until free calculation backend is found, then Computing and finally Completed.

In some cases, calculation task may be Failed due to bad (blurry/dark/etc.) photo was submitted or for some other reasons. In rare cases task may be Timed out - calculation took too long time.

It is your responsibility to poll the avatar calculation task for progress (status and progress fields of avatar DTO). Access to the URL fields of the incomplete avatar will not succeed.

Request parameters

Accepts multipart/form-data requests with the following fields:

• name - new avatar name
• description - new avatar description
• photo - new avatar source photo. Either photo or selfie must be provided. Error is returned if none or both are provided
• selfie - selfie code. Please see selfies section to learn more about selfies. Either selfie or photo must be provided. Error is returned if none or both are provided. Error is returned if selfie file was not uploaded yet by end-user
• pipeline - an optional field to specify which pipeline to use for new avatar generation. Please see available pipelines for the list of possible field values. Assumed value is animated_face if form field is not provided
• pipeline_subtype - modification of pipeline. Please see available pipelines for the list of possible field values
• parameters - specific parameters to compute for the submitted avatar. See parameters DTO for this field format. If no value submitted - replaced with default parameters for specified pipeline and (optionally) pipeline subtype. See parameters retrieval request to find out default/available parameters for pipeline type/subtype pairs. If no parameters required to compute avatar - submit empty JSON object: {}
• export_parameters - parameters to specify what resulting avatar downloadable data should include: like model format, set of textures, blendshapes, whether to embed all these assets into the one model or save as separate files, etc. Please see the export parameters section to find out more. Please note: if this section is empty for head_1.2 or head_2.0 pipelines - no avatar exports will be created.

4.4.4   Update

Method

PATCH, PUT

URL

/avatars/{code}/

Description

Update name/description of the particular avatar model identified by the code URL part. PATCH method supports partial update (e.g. only one field may be submitted), PUT method requires both fields to be submitted. Available in developer and client access modes. In client access mode restricted to avatars created by specified PlayerUID.

Request parameters

Accepts multipart/form-data requests with the following fields:

• name - new avatar name
• description - new avatar description

4.4.5   Delete

Method

DELETE

URL

/avatars/{code}/

Description

Delete the particular avatar model identified by code URL part. Available in developer and client access modes. In client access mode restricted to avatars created by specified PlayerUID.

4.5.1   Retrieve

Method

GET

URL

/avatars/count/

Description

Retrieve number of avatars created (created field), number of free avatars left (i.e. included into your billing plan, prepaid field) and number of avatars to be paid (to_pay field) in the current billing period. Please see the example response below:

{
  "created": 8192,
  "prepaid": 0,
  "to_pay": 2192
}

4.6.1   List

Method

GET

URL

/avatars/{code}/haircuts/

Description

Retrieve a list of per-avatar haircuts. The particular avatar is identified by the code URL part. Available in developer and client access modes. In client access mode restricted to avatars created by specified PlayerUID. Please see the avatar haircut DTO for the response format description.

Please note: you should not use this method to retrieve list of avatar haircuts for avatars created with export parameters: avatar exports list request should be used instead.

Please see the example response below:

[
  {
    "identity":"female_NewSea_J086f",
    "gender":"female",
    "preview":"https://api.avatarsdk.com/avatars/b63f9737-8594-460a-99a4-74e60b042bd4/haircuts/female_NewSea_J086f/preview/",
    "url":"https://api.avatarsdk.com/avatars/b63f9737-8594-460a-99a4-74e60b042bd4/haircuts/female_NewSea_J086f/",
    "texture":"https://api.avatarsdk.com/avatars/b63f9737-8594-460a-99a4-74e60b042bd4/haircuts/female_NewSea_J086f/texture/",
    "model":"https://api.avatarsdk.com/avatars/b63f9737-8594-460a-99a4-74e60b042bd4/",
    "pointcloud":"https://api.avatarsdk.com/avatars/b63f9737-8594-460a-99a4-74e60b042bd4/haircuts/female_NewSea_J086f/pointcloud/",
    "mesh":"https://api.avatarsdk.com/avatars/b63f9737-8594-460a-99a4-74e60b042bd4/haircuts/female_NewSea_J086f/mesh/"
  },
  ...
]

Filters

• gender - filter by haircut gender

4.6.2   Retrieve

Method

GET

URL

/avatars/{code}/haircuts/{identity}/

Description

Retrieve a particular instance of per-avatar haircut. Avatar is identified by code URL part, haircut - by identity URL part. Available in developer and client access modes. In client access mode restricted to avatars created by specified PlayerUID. Please see avatar haircut DTO for response object description.

Please note: you should not use this method to retrieve avatar haircuts for avatars created with export parameters: avatar export retrieval request should be used instead.

4.7.1   Retrieve

Method

GET

URL

/avatars/{code}/mesh/

Description

Retrieve zip archive with avatar mesh.

Please note: you should not use this method to retrieve avatar mesh for avatars created with export parameters: avatar export retrieval request should be used instead.

Avatars computed with animated_face, head_1.2 and head_2.0 (head/mobile and bust/mobile subtypes only) pipelines support different levels of mesh details (LOD). Specify desired LOD on mesh retrieval as a request parameter. Please find more information on mesh properties in the LOD table for each particular pipeline description in available pipelines section.

Request Parameters

• lod - level of details, integer in range [0, 8]

4.8.1   List

Method

GET

URL

/avatars/{code}/textures/

Description

Additional textures (along with default one) can be requested to compute for a particular avatar. See additional textures parameters to find out which textures can be requested.

Use this request to retrieve a list of available textures for the particular avatar. The particular avatar is identified by the code URL part.

Please see the example response below:

[
    {
        "file": "https://api.avatarsdk.com/avatars/33639df4-7593-44ab-a793-5e61a82bc81b/textures/slightly_cartoonish_texture/file/",
        "identity": "slightly_cartoonish_texture"
    },
    {
        "file": "https://api.avatarsdk.com/avatars/33639df4-7593-44ab-a793-5e61a82bc81b/textures/model_cartoonish/file/",
        "identity": "model_cartoonish"
    },
    {
        "file": "https://api.avatarsdk.com/avatars/33639df4-7593-44ab-a793-5e61a82bc81b/textures/model/file/",
        "identity": "model"
    }
]

4.8.2   Retrieve

Method

GET

URL

/avatars/{code}/textures/{identity}/file/

Description

Retrieve a particular texture file identified by its identity for particular avatar identified by code. Texture format may vary (i.e. JPEG, PNG, etc.), thus consult Content-Type HTTP Response Header for particular texture format.

4.9.1   Retrieve

Method

GET

URL

/avatars/{code}/blendshapes/

Description

Retrieve zip archive with all available blendshapes for the specific avatar. If no blendshapes available - the request will end with HTTP 204 NO CONTENT response.

Please note: you should not use this method to retrieve avatar blendshapes for avatars created with export parameters: avatar export retrieval request should be used instead.

By default, blendshapes are delivered in ply files packed into a zip archive. Blendshapes format may be changed to one fbx file also packed into a zip archive.

Avatars computed with animated_face, head_1.2 and head_2.0 (head/mobile subtype only) pipelines support different levels of blendshapes details (LOD). For more information on LOD details see avatar mesh retrieve request. Specify desired LOD on blendshapes retrieval as a request parameter.

Request Parameters

• fmt - blendshapes format. Available choices are: ply and fbx
• lod - level of details, integer in range [0, 8]

4.10.1   Retrieve

Method

GET

URL

/avatars/{code}/thumbnail/

Description

Retrieve thumbnail of submitted for calculations avatar photo.

Request Parameters

• max_w - limit thumbnail width, in range (0, 1024]
• max_h - limit thumbnail height, in range (0, 1024]

4.11.1   Retrieve

Method

GET

URL

/avatars/{code}/model_info/

Description

Retrieve JSON object with avatar model meta-information requested on avatar creation. Please see the model info parameters section for available parameters.

Please see the example response below:

{
    "age": "not_child",
    "age_confidence": 1,
    "gender": "male",
    "gender_confidence": 0.9999961853027344,
    "hair_color": {
        "blue": 114,
        "green": 145,
        "red": 199
    },
    "pipeline": "animated_face",
    "pipeline_subtype": "base/legacy",
    "skin_color": {
        "blue": 151,
        "green": 181,
        "red": 235
    }
}

4.12.1   Create

Method

POST

URL

/selfies/

Description

Create a new selfie upload page. Please see selfie DTO for response format.

4.12.2   retrieve QR code

Method

GET

URL

/selfies/{code}/qr/

Description

Retrieve QR code with upload page URL as PNG image file.

4.12.3   Retrieve

Method

GET, HEAD

URL

/selfies/{code}/file/

Description

Retrieve end-user photo. If the end-user has not uploaded the file yet, the request will end with HTTP 204 NO CONTENT. It is convenient to check file status with HEAD request first and download it with GET request once HTTP 200 response code returned.

4.13.1   List

Method

GET

URL

/avatars/{code}/exports/

Description

Retrieve a list of all available Avatar exports for a particular Avatar. The code key identifies the particular Avatar export. Please see the example response below:

[
    {
        "avatar_code": "e56d6044-fdbf-4cbf-86df-793bf21494f4",
        "code": "c0fc0ece-e291-4bbe-bdf2-ece270e0e065",
        "created_on": "2020-11-16T12:57:03.641470Z",
        "files": [
            {
                "category": null,
                "file": "https://api.avatarsdk.com/avatars/e56d6044-fdbf-4cbf-86df-793bf21494f4/exports/c0fc0ece-e291-4bbe-bdf2-ece270e0e065/files/avatar/file/",
                "identity": "avatar",
                "static_files": []
            },
            {
                "category": "haircuts",
                "file": "https://api.avatarsdk.com/avatars/e56d6044-fdbf-4cbf-86df-793bf21494f4/exports/c0fc0ece-e291-4bbe-bdf2-ece270e0e065/files/generated/file/",
                "identity": "generated",
                "static_files": []
            },
            {
                "category": "blendshapes",
                "file": "https://api.avatarsdk.com/avatars/e56d6044-fdbf-4cbf-86df-793bf21494f4/exports/c0fc0ece-e291-4bbe-bdf2-ece270e0e065/files/mobile_51/file/",
                "identity": "mobile_51",
                "static_files": []
            },
            {
                "category": "haircuts",
                "file": "https://api.avatarsdk.com/avatars/e56d6044-fdbf-4cbf-86df-793bf21494f4/exports/c0fc0ece-e291-4bbe-bdf2-ece270e0e065/files/afro/file/",
                "identity": "afro",
                "static_files": [
                    "https://api.avatarsdk.com/avatars/e56d6044-fdbf-4cbf-86df-793bf21494f4/exports/c0fc0ece-e291-4bbe-bdf2-ece270e0e065/files/afro/static_files/texture.png"
                ]
            }
        ],
        "status": "Completed",
        "url": "https://api.avatarsdk.com/avatars/e56d6044-fdbf-4cbf-86df-793bf21494f4/exports/c0fc0ece-e291-4bbe-bdf2-ece270e0e065/"
    }
]

4.13.2   Retrieve

Method

GET

URL

/avatars/{code}/exports/{export_code}/files/{export_file}/file/

Description

Retrieve particular Avatar export file. Avatar is identified by code URL part, Avatar export is identified by export_code URL part and export file is identified by export_file URL part. Please see avatar export DTO to see where to find each URL part.

Please note: all export files and their static files must be downloaded to retrieve full Avatar export.

Please see also note on static export files. Static file cache key may be constructed by concatenating Avatar export file category, identity, and static filename. E.g., for the texture of afro haircut from the example above, cache key may be: haircuts/afro/texture.png