NAV
shell

Introduction

Welcome to the FoxyAI API!

FoxyAI provides a simple and powerful REST API to integrate Visual Property Intelligence tools into your business or application.

The FoxyAI team loves to hear your comments and feedback. Email us! support@foxyai.com

Authentication

To authorize, use this code:

# With shell, you can just pass the correct header with each request
curl "api_endpoint_here" \
  -H "Authorization: API_KEY"

Make sure to replace API_KEY with your API key.

All requests are authenticated using an API Key provided by the FoxyAI team. To request a key, send an email to support@foxyai.com.

Header Value
Authorization Bearer {API_KEY}
Content-Type application/json

Image Analysis

Analyze Images

curl -X "POST" "https://api.remodelit.ai/property-image/batch" \
     -H 'authorization: bearer {api_key}' \
     -H 'Content-Type: application/json' \
     -d $'{
            "urls": [
              "https://photos.zillowstatic.com/fp/7964d9c3c45de20cea8e515ce3877d76-uncropped_scaled_within_1536_1152.jpg",
              ...
            ]
          }'

Response:

{
  "metrics": {
    "Kitchen": {
      "count": 1,
      "average_condition": 2.49722075
    },
    "Overall": {
      "count": 1,
      "average_condition": 2.49722075
    }
  },
  "images": [
    {
      "class": {
        "status": "complete",
        "type": "Kitchen",
        "confidence": 0.998914361
      },
      "quality": {
        "status": "complete",
        "score": 2.49722075
      },
      "objects": {
        "status": "pending",
        "items": []
      },
      "url": "https://photos.zillowstatic.com/fp/7964d9c3c45de20cea8e515ce3877d76-uncropped_scaled_within_1536_1152.jpg",
      "createdAt": "2021-03-26T17:38:49.020Z",
      "updatedAt": "2021-03-26T17:38:49.021Z",
      "id": "seJamnCgDuEpLXIw"
    }
    ...
  ]
}

The image analysis endpoint allows you to post a list of image urls and receive inference about those images.

HTTP Request

Method Path
POST /property-image/batch

Query Parameters

Parameter Required Description
urls yes Array of image URLs to be analyzed.
objectDetection no If set to true, this will enable object detection for the images passed in urls parameter. See below for more information on retrieving the results of object detection.

Object / Material Detection

Step 1: Post image(s)

curl -X "POST" "https://api.remodelit.ai/property-image/batch" \
     -H 'authorization: bearer {api_key}' \
     -H 'Content-Type: application/json' \
     -d $'{
            "objectDetection": "true",
            "urls": [
              "https://photos.zillowstatic.com/fp/7964d9c3c45de20cea8e515ce3877d76-uncropped_scaled_within_1536_1152.jpg"
            ]
          }'

Step 2: Get object detection results by image ID (returned in Step 1 response)

curl -X "GET" "https://api.remodelit.ai/property-image/abcdef" \
     -H 'authorization: bearer {api_key}' \
     -H 'Content-Type: application/json'

Response:

{
  ...
  "images": [
    {
      ...
      "objects": {
        "status": "complete",
        "items": [
          {
            "centroid": {
              "x": 0.22,
              "y": 0.8208232445520581
            },
            "class": "Floor",
            "materials": [
              {
                "type": "carpet",
                "confidence": 0.7804508019072388
              },
              {
                "type": "fabric",
                "confidence": 0.16634156913740789
              },
              {
                "type": "stone",
                "confidence": 0.04789770264412657
              },
              {
                "type": "brick",
                "confidence": 0.005309926311226701
              }
            ]
          }
        ]
      }
    }
  ]
}

Object detection can be enabled by passing objectDetection: true in the post body of the image analysis endpoint. This will queue object detection to be run asynchronously and you must poll the GET /property-image/{id} endpoint to receive the results.

The following objects classes will be detected and labeled: ["Chimney", "Building", "Wall", "Ceiling", "Window", "Dishwasher", "Sink", "House", "Fence", "Railing", "Stairs", "Toilet", "Roof", "Microwave", "Shower", "Bathtub", "Fireplace", "Skylight", "Fountain", "Area Rug", "Kitchen Island", "Washer/Dryer", "Hood", "Stove/Range/Oven", "Refrigerator", "Cabinets", "Countertops", "Double Door", "Ceiling Light", "Floor", "Chandelier"]

HTTP Request

Method Path
GET /property-image/{id}

Query Parameters

Parameter Required Description
urls yes Array of image URLs to be analyzed.
objectDetection no If set to true, this will enable object detection for the images passed in urls parameter. See below for more information on retrieving the results of object detection.

Flywheel (Image Feedback)

curl -X "POST" "https://api.remodelit.ai/flywheel" \
     -H 'authorization: bearer {api_key}' \
     -H 'Content-Type: application/json' \
     -d $'{
            "image": "{image_id_from_batch}",
            "class": {
              "confidence": 0.999325395,
              "type": "Frontyard"
            },
            "quality": {
              "score": 1.33736598
            }
          }'

Response:

{
  "image": "{image_id_from_batch}",
  "url": "https://remodelit-media.s3.amazonaws.com/image.png",
  "key": "{your_api_key}",
  "class": {
    "type": "Frontyard",
    "confidence": 0.999325395
  },
  "quality": {
    "score": 1.33736598
  },
  "createdAt": "2022-01-05T00:42:32.806Z",
  "updatedAt": "2022-01-05T00:42:32.806Z",
  "id": "0o0yEGMFk9FxivHo"
}

The Flywheel endpoint allows you to send us corrected prediction results which we can use to tune our models.

HTTP Request

Method Path
POST /flywheel

Query Parameters

Parameter Required Description
image yes The image id you are sending corrections for.
class.type no The corrected class type of the image
class.confidence no The corrected class confidence of the image (Most likely 1)
quality.score no The corrected quality score of the image

Renovation

Renovation Cost

curl -X "POST" "https://api.remodelit.ai/renovation-cost" \
     -H 'authorization: {api_key}' \
     -H 'Content-Type: application/json; charset=utf-8' \
     -d $'{
            "zipCode": "11201",
            "projectTypes": [
              "refinishWoodFloors",
              "installNewCarpet"
            ]
          }'

Response:

{
  "installNewCarpet": {
    "cost": {
      "range": [679, 2206],
      "avg": 1443,
      "min": 220,
      "max": 3850
    },
    "projectType": "installNewCarpet",
    "zipCode": "11201",
    "createdAt": "2021-04-19T16:39:15.580Z",
    "projectName": "Install New Carpet",
    "updatedAt": "2021-04-19T16:39:15.580Z",
    "id": "Qi2d1Xr2GfkDe1lS"
  },
  "refinishWoodFloors": {
    "cost": {
      "range": [1067, 2470],
      "avg": 1744,
      "min": 550,
      "max": 3960
    },
    "projectType": "refinishWoodFloors",
    "zipCode": "11201",
    "createdAt": "2021-04-19T16:39:15.525Z",
    "projectName": "Refinish Wood Floors",
    "updatedAt": "2021-04-19T16:39:15.525Z",
    "id": "b2IwjEG5ieE0JgNq"
  }
}

The renovation cost API returns pricing data for common renovation projects by zip code.

HTTP Request

Method Path
POST /renovation-cost

Query Parameters

Parameter Required Description
zipCode yes Zip code of project to be analyzed.
projectTypes yes Array of project type keys to be analyzed. For a list of available project types and their corresponding keys, leverage the /renovation-cost/project-type endpoint below.

Project Types

curl -X "GET" "https://api.remodelit.ai/renovation-cost/project-types" \
     -H 'authorization: bearer {api_key}' \
     -H 'Content-Type: application/json'

Response:

[
  {
    "key": "refinishWoodFloors",
    "name": "Refinish Wood Floors"
  },
  {
    "key": "installNewCarpet",
    "name": "Install New Carpet"
  },
  {
    "key": "replaceCarpetWithHardwood",
    "name": "Replace Carpet with Hardwood"
  },
  ...
]

The project-types endpoint returns a list of project types available for analysis.

HTTP Request

Method Path
POST /renovation-cost/project-types

Query Parameters

N/A

Property Preservation

Predict

curl -X "POST" "https://api.remodelit.ai/guardian/predict" \
     -H 'Content-Type: application/json; charset=utf-8' \
     -d $'{
  "url": "https://remodelit-media.imgix.net/2021/11/01/fceae94d-908e-49cf-a0a4-ff18b4fe5965.jpg",
  "model": "streetSigns"
}'

Response (accepted):

{
  "result": [
    [
      0.3181952238082886, 0.4270417392253876, 0.6492163658142089,
      0.5270580649375916, 0.8875733017921448
    ]
  ]
}

Response (rejected):

{
  "result": []
}

The property-preservation endpoint analyzes standard images taken by property preservation inspectors and verifies that the submitted photos contain the desired object.

Objects include: street signs, house numbers, vacancy forms, and gps units.

If the image submitted is rejected, an empty array is returned. If the image is accepted, an array of 5 or 6 element subarrays is returned, one subarray for each object detected in the image. The 6 values in the subarrays are:

[x, y, xx, yy, probability, class]

Value Description
x The x-coordinate (percentage) of the top-left corner of the bounding box
y The y-coordinate (percentage) of the top-left corner of the bounding box
xx The x-coordinate (percentage) of the bottom-right corner of the bounding box
yy The y-coordinate (percentage) of the bottom-right corner of the bounding box
probabilty The confidence of the prediction
class Class name of detected object if using 'multiObjectDetection' model. Values: 'forms', 'gps', 'housenumbers' 'streetsigns', 'boarding', 'forsalesign', 'lockbox'

HTTP Request

Method Path
POST /guardian/predict

Query Parameters

Parameter Required Description
url yes Url of image to be analyzed.
model yes The model to be used to analyze the submitted image. Options include: gps, streetSigns, houseNumbers, vacancyForms, and multiObjectDetection

Batch Predict

Step 1: Post image(s)

curl -X "POST" "https://api.remodelit.ai/guardian/predict/batch" \
     -H 'authorization: bearer {api_key}' \
     -H 'Content-Type: application/json' \
     -d $'{
            "model": "multiObjectDetection",
            "urls": [
              "https://remodelit-media.imgix.net/2021/11/01/fceae94d-908e-49cf-a0a4-ff18b4fe5965.jpg",
              "https://remodelit-media.imgix.net/2021/11/02/d7deb430-8430-4753-ad8b-4a9a1df21b90.jpg"
            ]
          }'

Response:

{
  "images": [
    {
      "status": "pending",
      "url": "https://remodelit-media.imgix.net/2021/11/02/d7deb430-8430-4753-ad8b-4a9a1df21b90.jpg",
      "id": "34YuiaaXciXvUP02",
      "model": "multiObjectDetection"
    },
    {
      "status": "pending",
      "url": "https://remodelit-media.imgix.net/2021/11/01/fceae94d-908e-49cf-a0a4-ff18b4fe5965.jpg",
      "id": "YoomDWWuAoXUeP0B",
      "model": "multiObjectDetection"
    }
  ]
}

HTTP Request

Method Path
POST guardian/predict/batch

Query Parameters

Parameter Required Description
model yes name of the model to run against
urls yes array of the urls to run model on

Step 2: Get prediction results by image ID (returned in Step 1 response)

curl -X "GET" "https://api.remodelit.ai/guardian/predict/YoomDWWuAoXUeP0B" \
     -H 'authorization: bearer {api_key}' \
     -H 'Content-Type: application/json'

Response:

{
    "status": "complete",
    "id": "YoomDWWuAoXUeP0B",
    "url": "https://remodelit-media.imgix.net/2021/11/01/fceae94d-908e-49cf-a0a4-ff18b4fe5965.jpg",
    "model": "multiObjectDetection",
    "result": [
      [
        0.3181952238082886, 0.4270417392253876, 0.6492163658142089,
        0.5270580649375916, 0.8875733017921448, "streetsigns"
      ]
    ]
}

This will generate a record for each image and run the model against each asynchronously. GET /predict/{id} endpoint to receive the results.

HTTP Request

Method Path
GET guardian/predict/{id}

Query Parameters

Parameter Required Description
id yes id of the image to look up

Media Upload

Upload

curl -X "POST" "https://api.remodelit.ai/media/upload" \
     -H 'authorization: bearer {api_key}' \
     -H 'Content-Type: application/json; charset=utf-8' \
     -d $'{
  "contentType": "image/jpeg",
  "extension": "jpg"
}'

Response:

{
  "key": "2021/11/02/d7deb430-8430-4753-ad8b-4a9a1df21b90.jpg",
  "bucket": "remodelit-media",
  "uploadUrl": "https://remodelit-media.s3-accelerate.amazonaws.com/2021/11/02/d7deb430-8430-4753-ad8b-4a9a1df21b90.jpg?AWSAccessKeyId=AKIAWONLJKP23MP7SYGD&Cache-Control=public%2C%20max-age%3D31536000&Content-Type=image%2Fjpeg&Expires=1636411336&Signature=eRtZqL6b2iVmM25eMGd9tLZ4R6U%3D&x-amz-acl=public-read",
  "downloadUrl": "https://remodelit-media.imgix.net/2021/11/02/d7deb430-8430-4753-ad8b-4a9a1df21b90.jpg"
}

The media/upload endpoint can be used to request an upload URL for uploading various types of media files. Using the uploadUrl provided in the response, you can upload an image or video file with a PUT http request. This file can then be accessed at the downloadUrl also provided in the response from media/upload, which will not expire if you did not provide an api-key.

If using the media/upload endpoint with an api-key, the downloadUrl provided will be a temporary presigned url that expires in one hour and the image will be stored in a private S3 bucket

HTTP Request

Method Path
POST /media-upload

Query Parameters

Parameter Required Description
contentType yes Content type of the media file being submitted.
extension no Extension to be applied to the downloadURL returned in the response (example: "jpg")

Errors

The Foxy AI API uses the following error codes:

Error Code Meaning
400 Bad Request -- Your request is invalid.
401 Unauthorized -- Your API key is wrong.
403 Forbidden -- The endpoint requested is hidden for administrators only.
404 Not Found -- The specified endpoint could not be found.
405 Method Not Allowed -- You tried to access an endpoint with an invalid method.
406 Not Acceptable -- You requested a format that isn't json.
410 Gone -- The endpoint requested has been removed from our servers.
429 Too Many Requests -- You're making too many requests! Slow down!
500 Internal Server Error -- We had a problem with our server. Try again later.
503 Service Unavailable -- We're temporarily offline for maintenance. Please try again later.
shell