Skip to content

Releases: AssemblyAI/assemblyai-node-sdk

4.7.0

26 Aug 13:57
c126ce0
Compare
Choose a tag to compare

What's Changed

  • Add language_confidence_threshold to Transcript, TranscriptParams, and TranscriptOptionalParams.

    The confidence threshold for the automatically detected language.
    An error will be returned if the language confidence is below this threshold.

  • Add language_confidence to Transcript

    The confidence score for the detected language, between 0.0 (low confidence) and 1.0 (high confidence)

Using these new fields you can determine the confidence of the language detection model (enable by setting language_detection to true), and fail the transcript if it doesn't meet your desired threshold.

Learn more about the new automatic language detection model and feature improvements on our blog.

Full Changelog: AssemblyAI/assemblyai-csharp-sdk@1.0.1...1.1.0

v4.6.1

17 Jul 10:15
18a2b66
Compare
Choose a tag to compare

What's Changed

  • Remove conformer-2 from SpeechModel union type.
  • Remove conformer-2 deprecation warning

Full Changelog: v4.6.0...v4.6.1

v4.6.0

10 Jul 15:05
ed72e4f
Compare
Choose a tag to compare

What's Changed

  • Add more TSDoc comments for RealtimeService documentation
  • Add new LeMUR models
  • Add TranscriptWebhookNotification which is a union of TranscriptReadyNotification or RedactedAudioNotification
  • Add RedactedAudioNotification which represents the body of the PII redacted audio webhook notification.

For more information, see CHANGELOG.

New Contributors

Full Changelog: v4.5.0...v4.6.0

v4.5.0

17 Jun 10:08
66b6fd7
Compare
Choose a tag to compare

What's changed

Added

  • You can now retrieve previous LeMUR responses using client.lemur.getResponse<LemurTask>("YOUR_REQUEST_ID").
  • LeMUR functions now return usage with the number of input_tokens and output_tokens.

v4.4.7

12 Jun 12:06
2f75167
Compare
Choose a tag to compare

What's changed

  • Rename TranscriptService.redactions function to TranscriptService.redactedAudio.
  • Add TranscriptService.redactedAudioFile function.
  • Add workerd export to fix cache issue with fetch on Cloudflare Workers.

4.4.6

10 Jun 10:43
fe04165
Compare
Choose a tag to compare
  • Fix Rollup exports so __SDK_VERSION__ is properly replaced with the version of the SDK.

v4.4.5

31 May 19:25
99736e3
Compare
Choose a tag to compare

What's Changed

  • Add new PiiPolicy enum values

Full Changelog: v4.4.4...v4.4.5

v4.4.4

30 May 21:03
21cc568
Compare
Choose a tag to compare

What's Changed

  • Add an export that only includes the Streaming STT code. You can use the export
    • by importing assemblyai/streaming,
    • or by loading the assemblyai.streaming.umd.js file, or assemblyai.streaming.umd.min.js file in a script-tag.
  • Add new EntityType enum values

Full Changelog: v4.4.3...v4.4.4

v4.4.3

15 May 16:21
78d1382
Compare
Choose a tag to compare

What's Changed

  • Add react-native exports that resolve to the browser version of the library.
  • Update dependencies
  • Update README.md
    • Add CDN instructions
  • Add CONTRIBUTING.md

Full Changelog: v4.4.2...v4.4.3

v4.4.2

03 May 15:10
3d8bd1f
Compare
Choose a tag to compare

What's Changed

Changed

  • Caching is disabled for all HTTP request made by the SDK
  • Accept data-URIs in client.files.upload(dataUri), client.transcripts.submit(audio: dataUri), client.transcripts.transcribe(audio: dataUri).
  • Change how the WebSocket libraries are imported for better compatibility across frameworks and runtimes.
    The library no longer relies on a internal #ws import, and instead compiles the imports into the dist bundles.
    Browser builds will use the native WebSocket, other builds will use the ws package.

Full Changelog: v4.4.1...v4.4.2