Developer Api

A flexible API powering custom voice analytics.

API AccessRead the Docs

VoiceBase’s Developer API offers application companies an open platform for scalable & secure voice analytics.

Custom Analytics

Get the results you want for the solutions you need, using our robust analytics toolkit.

Flexible Integrations

Build integrations to wherever your data is and wherever you want your results to end up.

PCI Certified Platform

Process your voice data through VoiceBase’s PCI Level 1 DSS Certified platform.

The Developer API enables SaaS applications to extract voice analytics across any audio, video, or text content. Our platform was built with flexibility in mind. Get the results you want, with full access to all your data. Process massive volumes and never worry about downtime or system breaches. 

How it works

The VoiceBase API is easy to use. Extract your data from anywhere and get a simple JSON output for your analytic results.

?php

$bearertoken = ‘your-token’;
$configuration = ‘your-json-configuration’;
$metadata = ‘your-json-metadata’;
$mediaUrl = ‘your-media-url’;

$url = “https://apis.voicebase.com/v3/media”;
$headers = array(
   ‘Authorization: Bearer ‘ . $bearertoken,
‘Content-Type: multipart/form-data’
);
set_time_limit(0);
$curl = curl_init();
$params = array(
“configuration” => $configuration,
“metadata” => $metadata,
“mediaUrl” => $mediaUrl
);
curl_setopt($curl, CURLOPT_POST, 1);
curl_setopt($curl, CURLOPT_HTTPHEADER, $headers);
curl_setopt($curl, CURLOPT_URL, $url);
curl_setopt($curl, CURLOPT_POSTFIELDS, $params);
curl_setopt($curl, CURLOPT_RETURNTRANSFER, true);
$response = curl_exec($curl);
$err = curl_error($curl);
curl_close($curl);
if ($err) {
 echo “cURL Error #:” . $err;
} else {
 echo $response;
}

?>

curl --location --request POST 
'https://apis.voicebase.com/v3/media' \

--header 'Content-Type: multipart/form-data' \
--header 'Authorization: Bearer $TOKEN' \
--form 'mediaUrl=https://www.example.com/sample.wav' \
--form 'configuration={
  "speechModel": { "language": "en-US" }}'

#python v3-postman-python-post-mediaUrl.py

import requests

url = “https://apis.voicebase.com/v3/media/”

media_file = ‘https://www.example.com/sample.wav’

payload= {‘mediaUrl’: media_file}
payload[‘configuration’] = (‘configuration.json’, ‘{“speechModel” : {
“language” : “en-US”} }’,’application/json’)

headers = {
    ‘Authorization’: “Bearer $TOKEN”
    }

response = requests.post(url, files=payload, headers=headers)

print(response.text)

<?php
$bearertoken = 'your-token';
$configuration = 'your-configuration';
$metadata = 'your-json-metadata';
$mediapath = '@sample.wav';

$url = "https://apis.voicebase.com/v3/media";
$headers = array(
  'Authorization: Bearer ' . $bearertoken,
  'Content-Type: multipart/form-data'
);
set_time_limit(0);
$curl = curl_init();
$params = array(
"configuration" => $configuration,
"metadata" => $metadata,
"media" => $mediapath
);
curl_setopt($curl, CURLOPT_POST, 1);
curl_setopt($curl, CURLOPT_HTTPHEADER, $headers);
curl_setopt($curl, CURLOPT_URL, $url);
curl_setopt($curl, CURLOPT_POSTFIELDS, $params);
curl_setopt($curl, CURLOPT_RETURNTRANSFER, true);
$response = curl_exec($curl);
$err = curl_error($curl);
curl_close($curl);
if ($err) {
 echo "cURL Error #:" . $err;
} else {
 echo $response;
}
?>
curl --location --request POST 'https://apis.voicebase.com/v3/media/' \
--header 'Authorization: Bearer $TOKEN' \
--form 'configuration={
   "speechModel":{
      "language":"en-US"
   }
}' \
--form 'media=@/Media/sample.wav'
#python v3-postman-python-post-multipart-file.py

import requests

url = "https://apis.voicebase.com/v3/media"

media_file = open('sample.wav','rb')

payload = {'media': ('sample.wav', media_file, 'audio/mpeg')}
payload['configuration'] = ('configuration.json', '{"speechModel" : { 
"language" : "en-US"} }','application/json')
headers = {
    'Authorization': "Bearer $TOKEN"
    }

response = requests.post(url, files=payload, headers=headers)

print(response.text)

Extract Voice Data From Anywhere

VoiceBase can ingest your audio, video, or text content from anywhere. Whether that be from your IVR system, social media messages, help platforms, or more.

Use the Language You Know

Use the language you already know and don’t worry about anything but uploading your files to get results.

Upload Using a Mediaurl or the File

Upload your voice data via mediaUrl or the actual voice file with a simple request across the most common coding languages.

Get JSON Results Fast

Your analytics will come out as a simple JSON format result. This open format allows for an easy understanding of your voice data for rapid insights. Get one response when your data has been uploaded, and another with the final transcripts.

Simple yet Powerful Transcripts

Get your transcript results with the accompanying metadata. This example shows a stereo speaker example of a transcript. When each speaker’s turn begins and ends is included, as well as who was speaking. 

{
  "_links": {
    "self": {
      "href": "/v3/media/1ef668ce-0feb-484e-95f1-797e446ca889"
    },
    "progress": {
      "href": "/v3/media/1ef668ce-0feb-484e-95f1-797e446ca889/progress"
    },
    "metadata": {
      "href": "/v3/media/1ef668ce-0feb-484e-95f1-797e446ca889/metadata"
    }
  },
  "formatVersion": "3.0.1",
  "mediaId": "1ef668ce-0feb-484e-95f1-797e446ca889",
  "status": "accepted",
  "dateCreated": "2017-10-13T20:48:04.494Z",
  "metadata": {},
  "mediaContentType": "audio/x-aiff",
  "length": 14965
}
{
  "mediaId": "bc14632d-e81b-4673-992d-5c5fb6573fb8",
  "status": "finished",
  "dateCreated": "2017-06-22T19:18:49Z",
  "dateFinished": "2017-06-22T19:19:27Z",
  "mediaContentType": "audio/x-wav",
  "length": 10031,
  "transcript": {
     "words":[
        {
           "p":1,
           "s":2200,
           "c":1.0,
           "e":2350,
           "w":"agent",
           "m":"turn"
        },
        {
           "p":2,
           "s":2200,
           "c":0.537,
           "e":2300,
           "w":"Hello"
        },
        {
          "p":3,
          "s":2300,
          "c":1.0,
          "e":2300,
          "w":".",
          "m":"punc"
        },
        {
           "p":4,
           "s":2400,
           "c":0.1,
           "e":2550,
           "w":"caller",
           "m":"turn"
        },
        {
           "p":5,
           "s":2400,
           "c":0.975,
           "e":2500,
           "w":"Hi"
        }
     ]
  }
}

Custom Voice Analytics

Full access to the power behind VoiceBase’s full stack of analytic features. Customize speech-to-text transcription, redact sensitive info, and categorize your calls in an optimal way for your business.

Speech-To-text

Get custom text transcripts that include stereo speaker ID, custom vocab, word-by-word timestamps, and PCI Redaction. 

PCI Redaction

Automatically detect and redact PCI/PII information from your voice data using VoiceBase’s PCI DSS Level 1 Certified platform.

Metrics

Dive into the metrics from your conversations such as overtalk, silence, and sentiment for a deep understanding of what is being said.

Categorization

Take advantage of dozens of categories from pre-built libraries or build custom categories that make sense for your business needs.

Predictions

Detect & predict complex behavior, like customer churn, across all of your voice data with high accuracy using machine learning.

Text Analytics

Analyze conversations chat, email, CRM, and support text data for a complete view of customer interactions.

Our Platform

Built with flexibility in mind.

Scalable

VoiceBase’s API was built on a hybrid public/private cloud architecture. Whether your application requires a thousand minutes/month of processing or a hundred million minutes, VoiceBase has you covered.

Secure

Enterprise-grade security has been a key feature of VoiceBase’s design from the very beginning. Every aspect of both the system and its hosting environment was designed to isolate and protect customer data.

Reliable

Our system is highly distributed, fully redundant, and self-healing. We deliver consistent uptime, with no downtime windows for maintenance or system upgrades. Rely on us to get the results you want with no issues.

Core Technology

The technology powering VoiceBase’s award-winning analytics

Speech Engine

The VoiceBase speech engine is based on a large vocabulary speech recognition (LVSR) platform that utilizes deep learning neural networks to best determine what was said. VoiceBase has developed features such as custom vocabulary, parallel processing, and number formatting in order to provide industry-leading accuracy and analytical results.

Analytics

In addition to the transcript, VoiceBase automatically surfaces keywords, key phrases, and topics from recordings using proprietary automatic speech recognition (ASR) algorithms and customer input. This enables businesses to identify trends, perform competitive analysis and better track each customer’s journey. VoiceBase also enables complex queries with a SQL-like Query language built for voice analytics.

Machine Learning

The API platform uses machine learning to perform predictive analysis of spoken information to automatically classify complex and subjective events. Businesses can utilize Predictive Insights to improve sale conversion rates, detect pre-churn, increase efficiency, and automate processes such as quality compliance measures.

Certifications

VoiceBase Developer API

The powerful API enabling companies to perform flexible and reliable analysis across large volumes of voice data.

Already Have Access to the API?

Access all of your analytics in one place with VoiceBase’s dev portal.