Version: 1.18.0 (latest)

API Requests for working with attempts

Request for getting a list of attempts

This request returns a paginated list of attempts (endeavors) filtered by external links. Filtering by the creation time range is supported. Results are sorted by creation time; the sort order can be changed.

Request schema:
GET /lrs/endeavor?sort_order={sort_order}&page={page_number}&page_size={page_size}

Query parameters

Parameter	Description
`sort_order`	Sort order by creation time. Possible values: `asc` (ascending), `desc` (descending). Required.
`page`	Page number to request. Minimum value: `1`. Required.
`page_size`	Number of items per page. Allowed range: 1 to 100. Required.
`external_link`	Filter by external link. If not specified, returns all attempts with `external_link` equal to `null`.
`start_date`	Lower bound of the creation time range (inclusive).
`end_date`	Upper bound of the creation time range (exclusive).

Request headers

Header	Description
`token`	Token of the `video-recorder` service. Stored in the `video-recorder-token` secret in the `./cfg/video-recorder.secrets.json` file.

Example request

curl -X 'GET' \
  'http://192.168.10.1/lrs/endeavor?sort_order=asc&page=1&page_size=1' \
  -H 'accept: application/json' \
  -H 'token: token'

Example response

{
  "page": 1,
  "page_size": 1,
  "total_count": 8,
  "endeavor_list": [
    {
      "id": "e4823024-d578-4285-9d3b-0471a4f720b2",
      "external_link": null,
      "content": [
        {
          "id": "5e77c40a-f45e-4a7e-ade8-8f676bc691f1",
          "files": [
            {
              "s3_link": "lr.videos:motion_control_video-26-07-06-12_38_10-no-parent-e4823-5e77c.webm",
              "raw_data_in_base64": null
            }
          ],
          "type": 1,
          "parent_id": null,
          "info": {
            "motion_control_result": [
              { "result": true, "pattern": "right" },
              { "result": true, "pattern": "left" },
              { "result": true, "pattern": "left" }
            ]
          },
          "exception_info": null,
          "creation_date": "2026-07-06T12:38:10.960546Z",
          "last_modified": "2026-07-06T12:38:10.960546Z"
        },
        {
          "id": "ac16bccc-478e-48ec-80e6-7cd0666a5e83",
          "files": [
            {
              "s3_link": "lr.ref-images:reference_frame-26-07-06-12_38_10-5e77c-e4823-ac16b.jpeg",
              "raw_data_in_base64": null
            }
          ],
          "type": 3,
          "parent_id": "5e77c40a-f45e-4a7e-ade8-8f676bc691f1",
          "info": {
            "angles": [0, -2],
            "image_info": {
              "quality": {
                "value": true,
                "failed_checks": null
              },
              "deepfake": {
                "confidence": 0.00021070241928100586
              },
              "liveness": {
                "value": "real",
                "confidence": 0.9773591756820679,
                "attack_type": "none",
                "attack_type_scores": {
                  "none": 0.9773591756820679,
                  "photo": 0.0008513810934558892,
                  "replay": 0.007360487079161384,
                  "2d_mask": 0.0008028665221645174,
                  "3d_mask": 0.0017056756156491595,
                  "regions": 0.011920414007501181
                }
              }
            },
            "frame_number": 128
          },
          "exception_info": null,
          "creation_date": "2026-07-06T12:38:10.960546Z",
          "last_modified": "2026-07-06T12:38:10.960546Z"
        },
        {
          "id": "ce79a10b-26c1-4ccb-9c21-6f5439569a77",
          "files": [],
          "type": 5,
          "parent_id": null,
          "info": {
            "verification_info": {
              "fa_r": 0,
              "fr_r": 0.6750852465629578,
              "score": 0.9856855273246765,
              "distance": 2339
            }
          },
          "exception_info": null,
          "creation_date": "2026-07-06T12:38:12.671648Z",
          "last_modified": "2026-07-06T12:38:12.671648Z"
        }
      ],
      "creation_date": "2026-07-06T12:37:58.219058Z",
      "last_modified": "2026-07-06T12:37:58.219058Z"
    }
  ]
}

The response contains the following fields:

Field	Description
`page`	Current page number (matches the one in the request).
`page_size`	Number of items per page (matches the request).
`total_count`	Total number of attempts matching the filter criteria.
`endeavor_list`	Array of attempts on the current page. For details on the attempt object, see the Attempt Object Structure section.

Request for getting a specific attempt by ID

This request returns an attempt by its ID. The attempt ID can be obtained from web component callbacks.

Request schema:
GET /lrs/endeavor/{endeavor_id}

Request headers

Header	Description
`token`	Token of the `video-recorder` service. Stored in the `video-recorder-token` secret in the `./cfg/video-recorder.secrets.json` file.

Example request

curl -X 'GET' \
  'http://192.168.10.1/lrs/endeavor/e4823024-d578-4285-9d3b-0471a4f720b2' \
  -H 'accept: application/json' \
  -H 'token: token'

Example response

{
  "id": "e4823024-d578-4285-9d3b-0471a4f720b2",
  "external_link": null,
  "content": [
    {
      "id": "5e77c40a-f45e-4a7e-ade8-8f676bc691f1",
      "files": [
        {
          "s3_link": "lr.videos:motion_control_video-26-07-06-12_38_10-no-parent-e4823-5e77c.webm",
          "raw_data_in_base64": null
        }
      ],
      "type": 1,
      "parent_id": null,
      "info": {
        "motion_control_result": [
          { "result": true, "pattern": "right" },
          { "result": true, "pattern": "left" },
          { "result": true, "pattern": "left" }
        ]
      },
      "exception_info": null,
      "creation_date": "2026-07-06T12:38:10.960546Z",
      "last_modified": "2026-07-06T12:38:10.960546Z"
    },
    {
      "id": "ac16bccc-478e-48ec-80e6-7cd0666a5e83",
      "files": [
        {
          "s3_link": "lr.ref-images:reference_frame-26-07-06-12_38_10-5e77c-e4823-ac16b.jpeg",
          "raw_data_in_base64": null
        }
      ],
      "type": 3,
      "parent_id": "5e77c40a-f45e-4a7e-ade8-8f676bc691f1",
      "info": {
        "angles": [0, -2],
        "image_info": {
          "quality": { "value": true, "failed_checks": null },
          "deepfake": { "confidence": 0.00021070241928100586 },
          "liveness": {
            "value": "real",
            "confidence": 0.9773591756820679,
            "attack_type": "none",
            "attack_type_scores": {
              "none": 0.9773591756820679,
              "photo": 0.0008513810934558892,
              "replay": 0.007360487079161384,
              "2d_mask": 0.0008028665221645174,
              "3d_mask": 0.0017056756156491595,
              "regions": 0.011920414007501181
            }
          }
        },
        "frame_number": 128
      },
      "exception_info": null,
      "creation_date": "2026-07-06T12:38:10.960546Z",
      "last_modified": "2026-07-06T12:38:10.960546Z"
    },
    {
      "id": "ce79a10b-26c1-4ccb-9c21-6f5439569a77",
      "files": [],
      "type": 5,
      "parent_id": null,
      "info": {
        "verification_info": {
          "fa_r": 0,
          "fr_r": 0.6750852465629578,
          "score": 0.9856855273246765,
          "distance": 2339
        }
      },
      "exception_info": null,
      "creation_date": "2026-07-06T12:38:12.671648Z",
      "last_modified": "2026-07-06T12:38:12.671648Z"
    }
  ],
  "creation_date": "2026-07-06T12:37:58.219058Z",
  "last_modified": "2026-07-06T12:37:58.219058Z"
}

Request for retrieving content by S3 link

This request returns binary content by an S3 link. The S3 link can be obtained from the attempt.

Request schema:
GET /lrs/get_content/{content_link}

Request headers

Header	Description
`token`	Token of the `video-recorder` service. Stored in the `video-recorder-token` secret in the `./cfg/video-recorder.secrets.json` file.

Example request

curl -X 'GET' \
  'http://192.168.10.1/lrs/get_content/lr.ref-images%3Areference_frame-26-07-06-12_38_10-5e77c-e4823-ac16b.jpeg' \
  -H 'accept: application/json' \
  -H 'token: token'

API requests for manual photo processing

This section describes requests that allow face detection on an image, building a biometric template, and comparing two such templates.

Request for face detection on an image

This request performs face detection on the provided photo and returns its coordinates and key points.

Request schema:
POST /face-detector-face-fitter/v2/process/image

Example request

curl -X 'POST' \
  'http://192.168.10.1/face-detector-face-fitter/v2/process/image' \
  -H 'accept: application/json' \
  -H 'Content-Type: multipart/form-data' \
  -F 'image=@MinimumIMG.jpeg;type=image/jpeg'

Request body parameters (multipart/form-data)

Parameter	Description
`image`	Image to process. Supported formats: `jpeg`, `png`. Required.

Example response

{
  "_image": {
    "blob": "/9j/4A...",
    "format": "IMAGE"
  },
  "objects": [
    {
      "keypoints": { ... },
      "pose": {
        "yaw": -8.746259689331055,
        "roll": -0.09920386970043182,
        "pitch": 2.89003586769104
      },
      "confidence": 0.8694550395011902,
      "id": 0,
      "bbox": [0.2, 0.21, 0.8466666666666667, 0.8433333333333334],
      "class": "face"
    }
  ]
}

Response structure

Field	Description
`_image`	Original image passed as input, in Base64 format.
`objects`	List of objects detected in the image with processing results. If empty, the detector did not find any face.
`objects[*].bbox`	Bounding box of the detected face, given in relative coordinates (normalised from 0 to 1) in order `[x1, y1, x2, y2]`. To convert to absolute pixel coordinates, multiply x‑coefficients by the original image width and y‑coefficients by its height.

Request for face detection and template building

This request detects a face in the provided photo and builds a biometric search template for it.

Request schema:
POST /face-detector-template-extractor/v2/process/image

Example request

curl -X 'POST' \
  'http://192.168.10.1/face-detector-template-extractor/v2/process/image' \
  -H 'accept: application/json' \
  -H 'Content-Type: multipart/form-data' \
  -F 'image=@MinimumIMG.jpeg;type=image/jpeg'

Request body parameters (multipart/form-data)

Parameter	Description
`image`	Image to process. Supported formats: `jpeg`, `png`. Required.

Example response

{
  "_image": {
    "blob": "/9j/4....",
    "format": "IMAGE"
  },
  "objects": [
    {
      "template": {
        "_face_template_extractor_1000_12": {
          "blob": "qjdp2UD...",
          "format": "NDARRAY",
          "dtype": "uint8",
          "shape": [296]
        }
      },
      "confidence": 0.8694550395011902,
      "id": 0,
      "bbox": [0.2, 0.21, 0.8466666666666667, 0.8433333333333334],
      "class": "face",
      "keypoints": { ... },
      "pose": { ... }
    }
  ]
}

Response structure

Field	Description
`_image`	Original image passed as input, in Base64 format.
`objects`	List of objects detected in the image with processing results. If empty, the detector did not find any face.
`objects[*].template`	Biometric template information, including its binary data in Base64 format.

Request for liveness detection on an image

This request performs liveness detection for a previously detected face.

Request schema:
POST /liveness-estimator/v2/process/sample

Example request

curl -X 'POST' \
  'http://192.168.10.1/liveness-estimator/v2/process/sample' \
  -H 'accept: application/json' \
  -H 'Content-Type: application/json' \
  -d '{
  "objects": [ ... ]
}'

Request body parameters (application/json)

The input is the result of a face detection or template building request. The request body must contain:

Field	Description
`_image`	Original image in Base64 format, passed to the previous handler.
`objects`	List of detected objects (must contain at least one face). If empty, the algorithm cannot perform liveness estimation.

Example request body

{
  "_image": {
    "blob": "/9j/4AAQSkZJRg...",
    "format": "IMAGE"
  },
  "objects": [
    {
      "keypoints": { ... },
      "pose": { ... },
      "confidence": 0.8694550395011902,
      "id": 0,
      "bbox": [ ... ],
      "class": "face"
    }
  ]
}

Example response

{
  "_image": {
    "blob": "/9j/4AAQSkZJ...",
    "format": "IMAGE"
  },
  "objects": [
    {
      "keypoints": { ... },
      "pose": { ... },
      "confidence": 0.8694550395011902,
      "id": 0,
      "bbox": [ ... ],
      "class": "face",
      "liveness": {
        "confidence": 0.9966132640838623,
        "value": "real",
        "attack_type": "none",
        "attack_type_scores": {
          "none": 0.9966132640838623,
          "replay": 0.000978723779512078,
          "photo": 0.0021829818897145916,
          "regions": 9.736091756044222e-05,
          "2d_mask": 9.789085341223873e-05,
          "3d_mask": 2.977847593834453e-05
        }
      }
    }
  ]
}

Response structure

The response returns the same _image object and an extended objects list. For each detected face, a liveness block is added:

Field	Description
`objects[*].liveness.value`	Liveness algorithm verdict: `"real"` – live person.
`objects[*].liveness.confidence`	Algorithm confidence that the image shows a live person. Higher is better.
`objects[*].liveness.attack_type`	Recognised attack type (`"none"` means no attack).
`objects[*].liveness.attack_type_scores`	Algorithm confidence for each attack type.

Request for template comparison

This request compares two biometric templates obtained from the Template Building Request.

Request schema:
POST /verify-matcher/v2/process/sample

Example request

curl -X 'POST' \
  'http://192.168.10.1/verify-matcher/v2/process/sample' \
  -H 'accept: application/json' \
  -H 'Content-Type: application/json' \
  -d '{
  "objects": [ ... ]
}'

Request body parameters (application/json)

The input is a list of two templates to compare. Order does not matter.

The request body contains the objects array – a list of objects with templates. Each object must contain:

Field	Description
`template`	Biometric template. Must be copied entirely from the `template` field of the template building API response.
`class`	Object type. Always must be `"face"`.

Example request body

{
  "objects": [
    {
      "template": {
        "_face_template_extractor_1000_12": {
          "blob": "qjdp...",
          "format": "NDARRAY",
          "dtype": "uint8",
          "shape": [296]
        }
      },
      "class": "face"
    },
    {
      "template": {
        "_face_template_extractor_1000_12": {
          "blob": "qjdp...",
          "format": "NDARRAY",
          "dtype": "uint8",
          "shape": [296]
        }
      },
      "class": "face"
    }
  ]
}

Example response

{
  "objects": [ ... ],
  "verification": {
    "distance": 0,
    "fa_r": 0,
    "fr_r": 0.9825811982154846,
    "score": 0.9773591756820679
  }
}

Response structure

Field	Description
`objects`	List of objects passed in the request (returned unchanged).
`verification`	Result of comparing the two templates.
`verification.score`	Algorithm confidence that the faces belong to the same person. Higher means higher probability of a match.

Attempt object structure

Attempt object fields

Field	Description
`id`	Unique attempt identifier.
`external_link`	External link associated with the attempt.
`content`	List of content items (processed binary data). Each item contains links to binary data and processing results.
`creation_date`	Attempt creation date and time.
`last_modified`	Last modification date and time.

Content item fields

Field	Description
`id`	Unique content item identifier.
`files`	List of S3 links to files. Contains links to video or images depending on the content type. Verification frames are not stored.
`type`	Content type: `1` – motion control video, `3` – reference frame, `5` – verification frame.
`info`	Biometric processing results. Format depends on the content type.
`exception_info`	Information about errors that occurred during content processing.

Content Type 1 – motion control video

The info field contains motion analysis results as a list of actions and their verdicts.

{
  "info": {
    "motion_control_result": [
      {
        "result": true,
        "pattern": "right"
      },
      {
        "result": true,
        "pattern": "left"
      },
      {
        "result": true,
        "pattern": "left"
      }
    ]
  }
}

Field	Description
`result`	Action execution verdict (`true` – performed, `false` – not performed).
`pattern`	Action name. Possible values: `up`, `left`, `right`, `closer`, `farther`.

Content Type 3 – reference frame

The info field contains results of biometric processing of the reference frame.

"info": {
  "angles": [0, -2],
  "image_info": {
    "quality": { "value": true, "failed_checks": null },
    "deepfake": { "confidence": 0.00021070241928100586 },
    "liveness": {
      "value": "real",
      "confidence": 0.9773591756820679,
      "attack_type": "none",
      "attack_type_scores": { ... }
    }
  },
  "frame_number": 128
}

Field	Description
`quality.value`	Quality assessment verdict (`true` – quality acceptable).
`quality.failed_checks`	List of failed quality checks (if any).
`deepfake.confidence`	Algorithm confidence in the presence of deepfake signs. Lower is better.
`liveness.value`	Liveness verdict: `"real"` – live person.
`liveness.confidence`	Algorithm confidence that the image shows a live person. Higher is better.
`liveness.attack_type`	Recognised attack type (`"none"` – no attack).
`liveness.attack_type_scores`	Algorithm confidence for each attack type.

Content Type 5 – verification frame

The info field contains the result of comparing the reference frame with the frame from the web component.

{
  "info": {
    "verification_info": {
      "fa_r": 0,
      "fr_r": 0.6750852465629578,
      "score": 0.9856855273246765,
      "distance": 2339
    }
  }
}

Field	Description
`score`	Algorithm confidence in face match. Higher means higher probability that the faces belong to the same person.

API Requests for working with attempts

Request for getting a list of attempts​

Query parameters​

Request headers​

Example request​

Example response​

Request for getting a specific attempt by ID​

Request headers​

Example request​

Example response​

Request for retrieving content by S3 link​

Request headers​

Example request​

API requests for manual photo processing

Request for face detection on an image​

Example request​

Request body parameters (multipart/form-data)​

Example response​

Response structure​

Request for face detection and template building​

Example request​

Request body parameters (multipart/form-data)​

Example response​

Response structure​

Request for liveness detection on an image​

Example request​

Request body parameters (application/json)​

Example request body​

Example response​

Response structure​

Request for template comparison​

Example request​

Request body parameters (application/json)​

Example request body​

Example response​

Response structure​

Attempt object structure

Attempt object fields​

Content item fields​

Content Type 1 – motion control video​

Content Type 3 – reference frame​

Content Type 5 – verification frame​

Request for getting a list of attempts

Query parameters

Request headers

Example request

Example response

Request for getting a specific attempt by ID

Request headers

Example request

Example response

Request for retrieving content by S3 link

Request headers

Example request

Request for face detection on an image

Example request

Request body parameters (multipart/form-data)

Example response

Response structure

Request for face detection and template building

Example request

Request body parameters (multipart/form-data)

Example response

Response structure

Request for liveness detection on an image

Example request

Request body parameters (application/json)

Example request body

Example response

Response structure

Request for template comparison

Example request

Request body parameters (application/json)

Example request body

Example response

Response structure

Attempt object fields

Content item fields

Content Type 1 – motion control video

Content Type 3 – reference frame

Content Type 5 – verification frame