AutoSolve – AudioSolve Setup with Google Speech and Wit.ai

The latest feature we have added to AutoSolve is AudioSolve. This will allow for the use of Speech APIs to solve audio challenges. AudioSolve is meant to be a backup solving method and will not have the capacity of solving solutions such as your OneClick browsers.

Before you begin to set these up, I would highly recommend reviewing our AutoSolve - Getting Started Guide. It will walk you through some of the basics of AutoSolve, as well as help you setup AutoSolve with your bots. 

Google Speech | Wit.ai | Errors & Tips

Configuring Google Cloud Speech-to-Text API

Google Speech-to-Text is a cloud API that allows for speech recognition to transcribe audio using Google's AI technologies. Their free tier allows for 60 minutes of Speech-to-Text per account before charging for conversions. To view pricing for Google Cloud Speech-to-Text API Click here.  If you'd like to learn how to create an account and utilize it with AudioSolve, you can use the following steps to do so:

1. To use the Cloud Speech-to-Text API, you will need to have a Google account. After logging into a Google head over to the Cloud Speech-to-Text API
2. At this point, please ensure that you have a project selected. If you do not have a GCP Project, you will need to create one now.
   
3. After you have selected the appropriate project, you will need to click Enable.
   
   
4. After you have enabled the API, it will bring you to the GCP Page on Cloud Speech-to-Text. Before you proceed, you will need to create an API Key from the API & Services Dashboard. 
   
5. From the API & Services Dashboard, you will need to click on Create Credentials and then API Key. 
6. Copy that API key and add it as a Third Party Solver Key to the AutoSolve dashboard.
   
   
7. From the dashboard under Third Party Solver API Keys, click on Add API Key.
   
8. Select Google Speech-to-Text from the Service dropdown set an easy to identify nickname for your API Key, such as the email address for your Google Cloud account. Then paste your API Key from the Google Speech API website into the API Key field and click Save.
   
   
9. After you have added the API Key, you can switch back to OneClick and use the Sync Dashboard button, which will automatically sync your Google Speech-to-Text API Key from the AutoSolve dashboard to OneClick.

Configuring OneClick

To use AudioSolve, you will need to have selected the corresponding API Key on the right. Once selected it will add itself to the list as a backup.

It is important to remember that Google Speech is a backup solving service. This particular service cannot be used without the use of OneClick Solvers. As such, you will not be able to connect it unless you have OneClick solvers started.

Again please note that Google Cloud Speech-to-Text API has pricing that is activated once you exit the free 60 minutes they have. Click here to view pricing. Make sure to view your billing dashboard in GCP regularly to ensure you are aware of all charges. 

Configuring Wit.ai

Wit.ai is a Facebook-owned solution in order to convert speech to text. In order to use it, you will need to sign into Wit.ai with a Facebook account, or a Github account. Unfortunately, the Github account integration will be deprecated and no longer usable come to the end of 2020.

1. Once you login, you will be on the Wit.ai dashboard. You will need to click the New App button.
   
2. After clicking on New App, you will be able to set a name for it, there are no requirements for the name, but it is always good to set it to something you can easily recognize. You can set the App to Private then press on Create.
   
3. After pressing create you will be on the Understanding tab of your app, from here, you will need to use the menu on the left in order to go to settings.
   
4. From settings, you will need to create a Client Access Token using the Generate new token button.
   
5. Once you generate a token, click it and copy it to your clipboard as this is the token you will need to connect to Wit.ai.
   
6. From the dashboard under Third Party Solver API Keys, click on Add API Key.
   
7. Select Wit.ai from the Service dropdown set an easy to identify nickname for your API Key, such as the app name. Then paste your Client Access Token from the Wit.ai website into the API Key field and click Save.
   
   
8. After you have added the Client Access Token to the Third Party API Key, you can switch back to OneClick and use the Sync Dashboard button, which will automatically sync your Wit.ai Access Token from the AutoSolve dashboard to OneClick.

Configuring OneClick

To use AudioSolve, you will need to have selected the corresponding API Key on the right. Once selected it will add itself to the list as a backup.

It is important to remember that Wit.ai is a backup solving service. This particular service cannot be used without the use of OneClick Solvers. As such, you will not be able to connect it unless you have OneClick solvers started.

Errors & Tips

When using AudioSolve, there is an error to be aware of, as well as some things to keep in mind to make the experience of using it that much better.

Errors

The error we will look at is the Try Again Later,  "Your Computer or network may be sending automated queries. To protect our users, we cannot process your request right now." This error occurs when Google has determined it has received too many Audio Challenge requests from your IP, or account. The error it will show is below.

When you get this error using AudioSolve, AutoSolve will reload the challenge and display the picture challenge, allowing you to solve it. Often if you click on the headphones for the Audio Challenge again, it will result in the same error requiring you to solve the picture challenge. 

Tips

One of the things we noticed while using AudioSolve, is that after using Audio Challenges to solve Captchas, ReCaptcha will default to Audio Challenges on that account when clicking the Checkbox. While we do not recommend using your personal Gmail account with AudioSolve but if you do, this is just something to keep in mind as it will default to the Audio Challenge anywhere you get a captcha.