Skip to content

Baidu AI Cloud Speech Recognition

Islet supports Baidu AI Cloud speech recognition. After setup, you can use it to transcribe voice recordings into text in Islet.

This feature uses your own Baidu AI Cloud account. Speech recognition requests are sent directly from Islet to Baidu AI Cloud, without passing through Islet servers.

Before you start

You need a Baidu AI Cloud account with real-name verification completed.

Step 1: Sign in or create an account

Open Baidu AI Cloud and sign in. If you do not have an account yet, follow the page prompts to create one.

Step 2: Complete real-name verification

After signing in, complete real-name verification in Baidu AI Cloud. Accounts without verification may not be able to create speech apps or use the speech recognition APIs.

Step 3: Open the speech app list

Open the speech app list.

If the page asks you to enable the speech service first, follow the prompts and then continue.

Step 4: Create an app

On the app list page, click "Create App" and fill in the form:

FieldWhat to enter
App nameAny name, for example "Islet Speech Recognition"
APIsSelect all APIs under speech technology
Speech package nameLeave empty
App ownershipPersonal
DescriptionAny description

Submit the form to create the app.

Step 5: Get the API Key and Secret Key

After the app is created, return to the speech app list.

Find the app you just created. Copy its API Key and Secret Key into Islet's speech recognition settings.

Keep your key private

Your Secret Key is a service credential. Do not share it publicly, and do not include it in screenshots, public docs, or chat messages.

Troubleshooting

  • Can't find the API Key or Secret Key? Make sure real-name verification is complete and the app was created successfully. Refresh the app list and check again.
  • Recognition fails in Islet? Check that the API Key and Secret Key were copied completely, and that speech technology APIs were selected when creating the app.
  • Do I need a speech package name? No. You can leave it empty for Islet's speech recognition setup.