wushenghua/hapi
1
0
Fork 0
mirror of https://github.com/wu736139669/hapi.git synced 2026-08-05 06:24:37 +00:00
Code Issues Packages Projects Releases Wiki Activity

Add realtime dictation providers (#1329)

Browse Source 
* feat: add realtime dictation providers

* fix: cancel realtime dictation startup

* fix: refresh local dictation availability

* fix: preserve dictation on disconnect

* fix: normalize OpenAI language hints
This commit is contained in:
SSU-WEI HUANG
2026-08-03 10:03:06 +08:00
committed by GitHub
parent 3c3bffdfbd
commit c3a5522207
21 changed files with 1378 additions and 46 deletions
Download Patch File Download Diff File Expand all files Collapse all files
shared/src/voice.backends.test.ts
+4 -2
View File
@@ -11,11 +11,13 @@ describe('listConfiguredTranscriptionProviders', () => {
expect(listConfiguredTranscriptionProviders({
OPENAI_API_KEY: 'openai',
ELEVENLABS_API_KEY: 'elevenlabs',
DEEPGRAM_API_KEY: 'deepgram',
TRANSCRIPTION_BASE_URL: 'http://localhost:8000/v1',
TRANSCRIPTION_MODEL: 'whisper-large-v3'
})).toEqual([
{ id: 'openai', label: 'OpenAI', modes: ['standard'] },
{ id: 'elevenlabs', label: 'ElevenLabs', modes: ['standard'] },
{ id: 'openai', label: 'OpenAI', modes: ['standard', 'realtime'] },
{ id: 'elevenlabs', label: 'ElevenLabs', modes: ['standard', 'realtime'] },
{ id: 'deepgram', label: 'Deepgram', modes: ['standard', 'realtime'] },
{ id: 'openai-compatible', label: 'OpenAI-compatible / local', modes: ['standard'] }
])
})
shared/src/voice.ts
+15 -5
View File
@@ -237,7 +237,14 @@ export type VoiceBackendType = 'elevenlabs' | 'gemini-live' | 'qwen-realtime'
export type VoiceMode = 'assistant' | 'dictation'
export type TranscriptionMode = 'standard' | 'realtime'
export type TranscriptionProvider = 'openai' | 'elevenlabs' | 'deepgram' | 'groq' | 'openai-compatible'
export type TranscriptionProvider = 'openai' | 'elevenlabs' | 'deepgram' | 'groq' | 'openai-compatible' | 'browser-local'
export const OPENAI_TRANSCRIPTION_MODEL = 'gpt-transcribe'
export const OPENAI_REALTIME_TRANSCRIPTION_MODEL = 'gpt-live-transcribe'
export const ELEVENLABS_TRANSCRIPTION_MODEL = 'scribe_v2'
export const ELEVENLABS_REALTIME_TRANSCRIPTION_MODEL = 'scribe_v2_realtime'
export const DEEPGRAM_TRANSCRIPTION_MODEL = 'nova-3'
export const GROQ_TRANSCRIPTION_MODEL = 'whisper-large-v3'
export interface TranscriptionProviderInfo {
id: TranscriptionProvider
@@ -246,13 +253,16 @@ export interface TranscriptionProviderInfo {
}
const TRANSCRIPTION_PROVIDERS: Record<TranscriptionProvider, TranscriptionProviderInfo> = {
openai: { id: 'openai', label: 'OpenAI', modes: ['standard'] },
elevenlabs: { id: 'elevenlabs', label: 'ElevenLabs', modes: ['standard'] },
deepgram: { id: 'deepgram', label: 'Deepgram', modes: ['standard'] },
openai: { id: 'openai', label: 'OpenAI', modes: ['standard', 'realtime'] },
elevenlabs: { id: 'elevenlabs', label: 'ElevenLabs', modes: ['standard', 'realtime'] },
deepgram: { id: 'deepgram', label: 'Deepgram', modes: ['standard', 'realtime'] },
groq: { id: 'groq', label: 'Groq', modes: ['standard'] },
'openai-compatible': { id: 'openai-compatible', label: 'OpenAI-compatible / local', modes: ['standard'] }
'openai-compatible': { id: 'openai-compatible', label: 'OpenAI-compatible / local', modes: ['standard'] },
'browser-local': { id: 'browser-local', label: 'Browser on-device', modes: ['realtime'] }
}
export const BROWSER_LOCAL_TRANSCRIPTION_PROVIDER = TRANSCRIPTION_PROVIDERS['browser-local']
/** Transcription providers whose startup environment is complete. */
export function listConfiguredTranscriptionProviders(env: VoiceBackendEnv): TranscriptionProviderInfo[] {
const providers: TranscriptionProvider[] = []