We are building a feature in our custom agent desktop app to display the full voice-to-text transcript of a call directly in the UI. The goal is to let agents review what was said without digging through the standard reporting interface. We have been using the Genesys Cloud Embeddable Client App SDK to handle authentication and then making direct HTTP requests to the Speech and Text Analytics API.
The issue is that when we call GET /api/v2/analytics/speech/transcripts/{conversationId}, the response only contains the first few sentences of the conversation for calls longer than 5 minutes. Short calls return the full text correctly. We have verified that the transcript is complete in the Genesys Cloud web UI, so the data exists in the backend.
Here is the code snippet we are using to fetch the data:
const response = await client.platformClientApi.speechAnalyticsApi.getTranscript(conversationId, null, null, null, null, 50);
console.log('Transcript length:', response.body.transcript.length);
We have tried increasing the limit parameter up to the maximum allowed value, but it does not change the result. The response status code is always 200 OK. We also noticed that the nextPageUri field in the response is null, even though we know there is more content. This suggests the API is not paginating the transcript content itself, but rather truncating it.
We have checked the API documentation and it mentions that transcripts can be large. Is there a different endpoint we should be using to retrieve the full text? Or is there a specific query parameter we are missing to force the API to return the complete transcript? We have also tried using the POST /api/v2/analytics/speech/transcripts/query endpoint with a specific conversation ID filter, but we get the same partial result.
We are using the latest version of the JavaScript SDK. Any help would be appreciated. The current implementation is blocking our release schedule.