No results found.
This page exists as a reference for old releases. New releases notes will be found on the Github page of the project.
v0.15.1 - 2021-11-05
- Add a
visibilityattribute to the generated API reference documentation.
v0.15.0 - 2021-10-28
- Add support for the following entities:
v0.14.1 - 2021-07-01
- Add JSDoc to
v0.14.0 - 2021-07-01
- Add support for
RelationLocationevents. Relation events group entities.
v0.13.1 - 2021-06-10
- Update the documentation generator.
- Update outdated dependencies.
v0.13.0 - 2021-05-25
- Fixed a bug with a regex replacing entities globally, and not only where is was supposed to do it.
- Added property
countryto the Uhlive constructor, which is used to format some entities. It will try to get the information from your browser and fallback to
modelproperty will now try to get the information from your browser and fallback to
onEntityFoundevents have the new
countryproperty to reflect which country formatting rules are used.
v0.12.0 - 2021-03-29
- Added support for event
EntityEvent.LocationCountry, which is triggered when a country is detected in the transcript.
- Added a confidence score property for the entities (
v0.11.0 - 2021-03-16
- It is now possible to listen to all
EntityFoundevents at once with the
Conversation.onEntityFound("*", () => ...)method.
v0.10.1 - 2021-03-10
- Added missing enums
- Added missing types for TypeScript users (
v0.10.0 - 2021-03-04
Conversation.onSpeakerLeft's callback now have as parameter a
SpeakerLeftobject instead of the speaker id (
- Some types have changed:
ConversationEventis removed in favor of
EntityEventto separate NER events for the others.
EntityOrdinalare replaced by
Entity, since all entities have the same signature.
- Event listeners
Conversation.onOrdinalFoundhave been replaced by the more generic
Conversation.onEntityFound, the first parameter being the entity we want to listen to.
speaker_joinedevent (see documentation).
- Support for Named Entity Recognition (NER) events (see documentation).
timestampproperty to speaker_left` event (see documentation).
joinmethod (see documentation).
v0.9.0 - 2021-02-01
- The default model is now
default, which was a model in French language. If you want to use a French model, please use
fr. You can see the list of available models here, and how to set a different model here.
- The library is now plug-and-play with very minimal configuration. Setup your identifier and token, and voilà, the transcription is automatically shown in a html tag. See more in the documentation.
- Improve documentation and returned types.
v0.8.2 - 2021-01-21
- Improve documentation
v0.8.1 - 2021-01-20
- Fix: Enrich events:
onEntityOrdinalFoundreturn the conversation.
v0.8.0 - 2021-01-11
- Support new "enrich" events:
onErrornew reports more information to help debug the error.
v0.7.0 - 2020-09-25
- When joining a converation, there is now three new possible parameters:
v0.6.0 - 2020-09-25
- Improve security, you now need to configure the
identifierparameter be identified.
v0.5.0 - 2020-09-17
- The SDK starts to stream the audio as soon as the user join a conversation in non-readonly mode
- Update NPM dependencies
- Better management of version numbers
Conversation.stopRecordinghave been removed since this in now done automaticaly when joining/leaving a conversation
v0.4.0 - 2020-08-28
- Display logs in the example
- Improve UX/UI of example
- Update outdated npm dependencies
v0.3.0 - 2020-08-13
- Support the new event
- Remove support for previous
transcriptevent in favor of
Table of content