Back to vidrovr.com/careers

Get started

Vidrovr provides full api solutions to embed video AI into your platform and products. In this tutotial we will be detailing two core components of our API: how send us your video and how to get the video AI results. Now let’s take a quick tour of how to get your first video running in the Vidrovr machine learning pipeline.

Getting your api key

Before you get started, you need to obtain a private api key from our system.

  • You can receive a log-in to our platform and API key directly by reaching out to our customer support team, or create an account for yourself on our dashboard page.
  • If you already have an account, you can login to vidrovr dashboard. The api key will be displayed on the homepage.
  • If you lost your api key, or your login credentials, feel free to email us at contact@vidrovr.com to retrieve your api key.

Upload content

Now, select a file that you’d like to test. We accept all major video file containers and encoders, you can see the full list on the video upload explainer API page.

Upload with restful api

You can use either your login credentials or api key to start the uploading. Here’s a sample of using your login credentials.

curl -F upload_type='video' -F  id=`USERNAME` -F password=`PASSWORD`
    -F filename=`FILENAME` -F data=`video_file.mp4`
    https://platform.vidrovr.com/upload/uploader

Sample response

{
 "msg":"you just have to play nice",
 "id":"c0e4641ed361f76e57bc679f383d9b49315ff97b-f889-4f3e-9efb-eafb7f00d43d",
 "resp":"upload_success"
}

Inside of this response:

  • id refers to a uniqle video asset id for the video that was uploaded to the Vidrovr system, this id can be used in metadata and asset retrieval across the whole vidrovr platform. This id is unique and will not be changed.

Get results

After you upload your content, Vidrovr’s backend engine will start processing your video. Depending on the size of your file, the processing time will vary. In general, a 30 minute long, standard hi-res (1080p) video will take us 10 minutes to prepare the data.

To get the metadata that we generated for your video, you can use a GET request on the URL below

https://platform.vidrovr.com/public/api/v01/get_metadata?id=&api_key=

Inside of this api call, the api_key is the private api key related to your account from Step 1. The id here refers to the video asset id you get after a successful content upload via our upload api in Step 2.

If the call returns a success, congratulations! You have just received a detailed understading of what is appearing in your video in the form of the Vidrovr metadata

Sample response

{
 "other_metadata":[...],
 "audio_words":[...],
 "name":"nat_geo_sharks_zone1.mp4",
 "tags":[...],
 "hashtags":[...],
 "scenes":[...],
 "id":"...",
 "keyphrases":[...],
 "on_screen_text":[...]
 "person_identification":[...],
 "creation_date":1500229993885,
 "thumbnail":"..."
}{
 "other_metadata":[...],
 "audio_words":[...],
 "name":"nat_geo_sharks_zone1.mp4",
 "tags":[...],
 "hashtags":[...],
 "scenes":[...],
 "id":"...",
 "keyphrases":[...],
 "on_screen_text":[...]
 "person_identification":[...],
 "creation_date":1500229993885,
 "thumbnail":"..."
}

Key properties in the response

Property
Type
Description
audio_words
array
This is the transcript generated from the audio within the video, each word that is detected comes with a confidence score and time of appearance.
name
string
The original file name provided during upload.
tags
tags
This contains a list of recognized concepts from your video. Each tag will have a tags property which represents the name of the tag, a start timecode, and a end timecode, and a confidence
scenes
array
This contains a list of recognized scenes from your video. Each scene will have a scenes property which represents the name of the scene, a start timecode, and a end timecode, and a confidence
on_screen_text
array
It includes a list of words that appears on screen. In addition to the standard start timecode and end timecode, the return object will also include a bounding box which details the position of this word on a 640 by 360 resolution video, these properties are x for x axis ,y for y axis, h for height, w for width
person_identification
array
This contains a list of recognized people from your video. Each person will have a person_name property which represents the name of the person, a start timecode, and a end timecode, and a confidence

More api calls

Our full available api list can be found on Vidrovr api page

On this site, we also go into detail around some of the concepts covered on this page.

  • Upload Content provides the details of how to use our restful upload method, HTTP Live Streaming (HLS)upload, and webhooks interface to upload content.
  • Get Results offers explanations of ways to get results back from the system. These methods include
    get_metadata which is the essential methods to get full metadata from a given video id.
    get_video_list which provides a list of the video in your account in chronological order.
    search offers a way to find videos in your account that are all relevant based on the metadata in your account.

Use our dashboard

Vidrovr also provides a free dashboard which provides a user interface for uploading content, inspect, searching through, and editing metadata, and to configure third party connections and custom model tasks.
For more information and instructions on how to use the dashboard. Please follow this link to learn more.

Understand your api limits

Your api calls are subject to operation limits, which can be seen in the dashboard. Feel free to reach out to us if you have overused your quota.