> ## Documentation Index
> Fetch the complete documentation index at: https://docs.bland.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# June 30, 2025

> Babel improvements, SSO, and voice over web chat

### Upgraded Babel Model + Language Locking

We’ve rolled out an upgrade to Babel, our in-house transcription engine—along with new language-locked variants for improved accuracy.

* Babel becomes more accurate over the course of the call using contextual speech refinement
* New API options support locked languages like English, Spanish, French, and more (e.g. `babel-en`, `babel-es`, `babel-fr`)
* Language locking improves recognition accuracy by removing the need for auto-detection
* Low-confidence audio is now marked with \[inaudible], allowing your agent to respond or prompt naturally

If your callers consistently speak the same language, we recommend trying a locked-language model for higher transcription quality.

<img src="https://mintcdn.com/blandai/J1ftAy4aMsUUPhgH/changelog_assets/06_30_2025/babel.jpg?fit=max&auto=format&n=J1ftAy4aMsUUPhgH&q=85&s=688163d70a6501f565df21e73c9c170e" style={{ width: "100%", borderRadius: "0.5rem", marginTop: "1rem", marginBottom: "1rem" }} alt="Babel Model" width="5088" height="3744" data-path="changelog_assets/06_30_2025/babel.jpg" />

***

### OIDC SSO Support (Enterprise)

We now support **OIDC-based Single Sign-On** for enterprise organizations.

* A new **SSO tab** is available in org settings for configuring OIDC providers
* The login flow has been updated to handle OIDC callback
* Works with identity platforms like Okta, Azure AD, and Google Workspace

<Tabs>
  <Tab title="Login Options">
    <img src="https://mintcdn.com/blandai/J1ftAy4aMsUUPhgH/changelog_assets/06_30_2025/SSO_login.jpeg?fit=max&auto=format&n=J1ftAy4aMsUUPhgH&q=85&s=3cb82f03cc34723c5cd9c02f4ebad428" style={{ width: "100%", borderRadius: "0.5rem", marginTop: "1rem", marginBottom: "1rem" }} alt="OIDC SSO" width="5088" height="3744" data-path="changelog_assets/06_30_2025/SSO_login.jpeg" />
  </Tab>

  <Tab title="SSO Login">
    <img src="https://mintcdn.com/blandai/J1ftAy4aMsUUPhgH/changelog_assets/06_30_2025/SSO_login_2.jpeg?fit=max&auto=format&n=J1ftAy4aMsUUPhgH&q=85&s=fee55af693099b0fbdbe5acedc845927" style={{ width: "100%", borderRadius: "0.5rem", marginTop: "1rem", marginBottom: "1rem" }} alt="OIDC SSO" width="5088" height="3744" data-path="changelog_assets/06_30_2025/SSO_login_2.jpeg" />
  </Tab>

  <Tab title="SSO Configuration">
    <img src="https://mintcdn.com/blandai/J1ftAy4aMsUUPhgH/changelog_assets/06_30_2025/SSO_configuration_bland.jpg?fit=max&auto=format&n=J1ftAy4aMsUUPhgH&q=85&s=f2950100be63df863822b9158e190eee" style={{ width: "100%", borderRadius: "0.5rem", marginTop: "1rem", marginBottom: "1rem" }} alt="OIDC SSO" width="5088" height="3744" data-path="changelog_assets/06_30_2025/SSO_configuration_bland.jpg" />
  </Tab>

  <Tab title="Add OIDC Provider">
    <img src="https://mintcdn.com/blandai/J1ftAy4aMsUUPhgH/changelog_assets/06_30_2025/SSO_configuration_modal.jpg?fit=max&auto=format&n=J1ftAy4aMsUUPhgH&q=85&s=c4cbaede5d05f4df0377f8bbf6d05b87" style={{ width: "100%", borderRadius: "0.5rem", marginTop: "1rem", marginBottom: "1rem" }} alt="OIDC SSO" width="5088" height="3744" data-path="changelog_assets/06_30_2025/SSO_configuration_modal.jpg" />
  </Tab>
</Tabs>

***

### Voice Chat in Web Widget

You can now enable **voice input** directly in the web chat widget—bringing the full conversational AI experience to the browser.

* Users can now talk with your agents, using the same speech pipeline as our phone calls.
* Perfect for embedded flows that need a faster, more natural interface
* Voice configuration available as an option in widget configuration
* Track all voice interactions through the Call Logs dashboard within Bland

<Tabs>
  <Tab title="Voice Chat">
    <img src="https://mintcdn.com/blandai/J1ftAy4aMsUUPhgH/changelog_assets/06_30_2025/web_widget_voice.jpeg?fit=max&auto=format&n=J1ftAy4aMsUUPhgH&q=85&s=d1e43916bc9f05f6eda0fb778f627958" style={{ width: "100%", borderRadius: "0.5rem", marginTop: "1rem", marginBottom: "1rem" }} alt="New Pathway Generator" width="5088" height="3744" data-path="changelog_assets/06_30_2025/web_widget_voice.jpeg" />
  </Tab>

  <Tab title="Voice Configuration">
    <img src="https://mintcdn.com/blandai/J1ftAy4aMsUUPhgH/changelog_assets/06_30_2025/web_widget_voice_active.jpeg?fit=max&auto=format&n=J1ftAy4aMsUUPhgH&q=85&s=865503da045d1dcfadcd095100cbe11f" style={{ width: "100%", borderRadius: "0.5rem", marginTop: "1rem", marginBottom: "1rem" }} alt="Pathway Editor" width="5088" height="3744" data-path="changelog_assets/06_30_2025/web_widget_voice_active.jpeg" />
  </Tab>
</Tabs>

***

### Improvements

* You can now configure a **"Go to Node"** action inside the **Press Button Node**
* Adds `isRetry: true` support in post call webhooks to determine if the call was retried
* Improved JSON validation when uploading conversational pathways to an organization
* Fixes test case issues affecting test coverage stability
* Fixes bug where if proxy agent presses button, primary call would go into end state (leave voicemail, or go to a node)
* Fixes a bug in pathway duplication logic
* Adds version tracking to pathway editor, so when a customer updates an old system of a pathway, it will error:

> *This pathway was updated elsewhere. Please refresh the page to get the latest changes.*

* Fixes a crash that could occur when referencing undefined variables in execution