fix: accumulate transcript chunks into single messages per turn
All checks were successful
Build & Push / build-and-push (push) Successful in 1m23s

Streaming transcription words now merge into one bubble per agent/user turn
instead of creating separate entries for each chunk.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
This commit is contained in:
2026-03-28 14:42:04 +01:00
parent c48313ad91
commit 66949c07d8

View File

@@ -132,6 +132,7 @@ export default function VoiceAgentProvider({ locale, children }: VoiceAgentProvi
const [completedBrief, setCompletedBrief] = useState<string | null>(null); const [completedBrief, setCompletedBrief] = useState<string | null>(null);
const [completedFormData, setCompletedFormData] = useState<WizardFormData | null>(null); const [completedFormData, setCompletedFormData] = useState<WizardFormData | null>(null);
const turnCompleteRef = useRef(true);
const wsRef = useRef<WebSocket | null>(null); const wsRef = useRef<WebSocket | null>(null);
const mediaStreamRef = useRef<MediaStream | null>(null); const mediaStreamRef = useRef<MediaStream | null>(null);
const audioContextRef = useRef<AudioContext | null>(null); const audioContextRef = useRef<AudioContext | null>(null);
@@ -141,7 +142,15 @@ export default function VoiceAgentProvider({ locale, children }: VoiceAgentProvi
const animFrameRef = useRef<number>(0); const animFrameRef = useRef<number>(0);
const addTranscript = useCallback((role: 'user' | 'agent', text: string) => { const addTranscript = useCallback((role: 'user' | 'agent', text: string) => {
setTranscript((prev) => [...prev, { role, text, timestamp: Date.now() }]); setTranscript((prev) => {
const last = prev[prev.length - 1];
// Append to last entry if same role and turn is still ongoing
if (last && last.role === role && !turnCompleteRef.current) {
return [...prev.slice(0, -1), { ...last, text: last.text + text }];
}
turnCompleteRef.current = false;
return [...prev, { role, text, timestamp: Date.now() }];
});
}, []); }, []);
const trackAmplitude = useCallback(() => { const trackAmplitude = useCallback(() => {
@@ -369,6 +378,10 @@ export default function VoiceAgentProvider({ locale, children }: VoiceAgentProvi
if (msg.serverContent.outputTranscription?.text) { if (msg.serverContent.outputTranscription?.text) {
addTranscript('agent', msg.serverContent.outputTranscription.text); addTranscript('agent', msg.serverContent.outputTranscription.text);
} }
// Turn complete — next output starts a new transcript entry
if (msg.serverContent.turnComplete || msg.serverContent.generationComplete) {
turnCompleteRef.current = true;
}
} }
// Tool call // Tool call