Invox Dictation SDK Version:

Changelog


Version 2.9.39 — November 2026

Breaking changes

  • INVOX.Login fails if the credential does not belong to organizationId. organizationId, appId and apiKey are all required and must belong to the same organization: associate the credential (appId and apiKey) with your organization in the portal. A credential without an organization, or associated with a different one, makes INVOX.Login fail with LOGIN_AUTHORIZATION_REJECTED, a code and text of its own, different from INVALID_CREDENTIALS.
  • Update the package to 2.9.39. Once the server check is deployed, versions earlier than 2.9.39 (including the 2.9.38 that the examples bundled before this release) stop working with the credentials of the Invox Dictation app, because they do not send the organization when requesting the token. From 2.9.39 the SDK always sends the organizationId of INVOX.init(...), and the credential of any app must be associated with that organization.

Version 2.9.38 — November 2026

Changed

  • INVOX.Login now rejects with an InvoxLoginError (translated message, code, httpStatus, exceptionType) instead of the internal client error, so the message can be shown to the user as is. See Login errors.

Version 2.9.1 — October 2026

Changed

  • transcriptionServiceUrl is no longer part of INVOX.init(...) and is rejected: the Direct Dictation endpoint is selected by the SDK build. It replaces the transcriptionServiceUrl option and its default host that version 2.9.0 introduced.

Version 2.9.0 — June 2026

Added

  • UI components (Bar, Dictionary, Templates, Transformations, Intelligent Formatting, Quality Review, Auto Fill Templates) available as named exports from the package.
  • Exported LoginCredentialsType in the public TypeScript API.

Changed

  • ESM and CommonJS bundles — removed the IIFE build. Use <script type="module"> or require() with a bundler.
  • Component CSS is auto-injected at runtime — no separate CSS file imports needed.
  • Opus audio library is bundled within the main module — no manual cp -r step required.
  • Added global SDK initialization via INVOX.init(...) for connection context.
  • INVOX.Login now accepts credentials only (user, password).
  • Connection context (transcriptionServiceUrl, adminServiceUrl, organizationId, appId, apiKey) is now provided through INVOX.init(...).
  • Added default init hosts when omitted: transcriptionServiceUrl defaults to wss://debug.invoxmedical.com and adminServiceUrl defaults to https://debug.invoxmedical.com/INVOX/.

Fixed

  • TypeScript declarations now emit correctly alongside the bundle.

Version 2.8.3.137 — February 24, 2025

Fixed

  • Fixed unnecessary line break included by CKEditor4 when the "New Paragraph" command was used.

Version 2.8.3 — October 11, 2024

Fixed

  • Minor problems corrected.

Version 2.8.1.4793 — September 18, 2023

Changed

  • Prevent loading Opus Audio Library in the head of the HTML if it has already been loaded in a previous session.

Fixed

  • Reduced the number of update events sent by the Dictation Service during the download of resources at login to avoid overloading the SDK.

Version 2.8.1.4779 — September 12, 2023

Fixed

  • Fixed a bug that prevented using multiple inline commands without pausing during dictation.

Version 2.8.1.4697 — August 30, 2023

Fixed

  • Fixed a bug that keeps the red icon in the browser tab after recording has stopped.

Version 2.8.1.4413 — August 14, 2023

Fixed

  • Fixed a bug that prevented the dictation start buttons from being used on handheld microphones.
  • Fixed a bug that increased the height of the dictation bar when templates contain newline characters.

Version 2.8.1.4359 — June 14, 2023

Added

  • Added JavaScript example in Quick Start and Dictation Bar Component sections about automating login for both Local and Remote Dictation Services.

Changed

  • Microphone names are now displayed with the full name set by the system.
  • Simplified and improved the explanations in the Quick Start section.

Fixed

  • Fixed audio noise issue by reverting to previous version of opus-recorder.
  • Fixed an issue where handheld microphones were using the speaker output as input in Remote Dictation Service scenarios.
  • Fixed a bug that caused a continuous loop of "Invox Dictation cannot access the microphone" message when no microphones were connected.
  • Fixed a bug in Firefox that prevented dictation after login.
  • Fixed a bug in Firefox that could not display the microphone name.
  • Fixed a bug in the dictation bar component whereby the status bar could overwrite an error message.

Version 2.8.1.3843 — March 16, 2023

Added

  • Added support for asynchronous writers.
  • Added code template for asynchronous writers in the documentation.
  • Added explanation in Dictation Bar Component about how to reset the component at runtime.
  • Added section about Session Management in the documentation.
  • Added section about Logging in the documentation.

Fixed

  • Fixed a bug in Writers that selects the first sentence if the cursor was already on the first sentence.

Version 2.8.1.3736 — March 3, 2023

Added

JavaScript SDK:

  • Added the non-minified script file of the Dictation Bar web component.
  • Added remove() public function to Dictation Bar, Dictionary, Templates, and Transformations web components.
  • Added ChangeLogKeywordsToIgnore function to the JavaScript API.
  • Added supported browsers in the documentation.
  • Added instructions to deploy examples in Docker in the documentation.

Recognition System:

  • Added internal functionality to filter identified noises.
  • Significant improvement in memory consumption at login.
  • Improvement in memory consumption by partial hypothesis.