ASR Providers

DeLive supports six ASR backends through a unified provider registry. Each provider implements a common contract but uses different transport and audio processing strategies.

Provider Comparison

Provider	Type	Transport	Audio	Streaming	Translation	Diarization
Soniox V4	Cloud	WebSocket	MediaRecorder (WebM/Opus)	Yes	Yes	Yes
Volcengine	Cloud	WebSocket (via proxy)	AudioWorklet (PCM16)	Yes	No	No
Groq	Cloud	REST (batch)	AudioWorklet (PCM16)	No	No	No
SiliconFlow	Cloud	REST (batch)	AudioWorklet (PCM16)	No	No	No
Local OpenAI	Local	REST (batch)	MediaRecorder (WebM/Opus)	No	No	No
whisper.cpp	Local	REST (local)	AudioWorklet (PCM16)	No	No	No

Execution Modes

Real-Time Streaming

Used by Soniox and Volcengine. Audio chunks are sent continuously over a WebSocket connection, and transcript updates arrive in real-time.

• Soniox emits token-level events (prefersTokenEvents: true) for fine-grained text updates
• Volcengine uses a local proxy (/ws/volc on port 23456) to inject required authentication headers

Windowed Batch

Used by Groq, SiliconFlow, Local OpenAI-compatible, and whisper.cpp. Audio accumulates in a rolling buffer (max 45 seconds), and a REST call retranscribes the entire window at regular intervals.

• Interval mode (Groq, SiliconFlow, whisper.cpp): retranscribe every 1.5 seconds
• Debounce mode (Local OpenAI): retranscribe 1200ms after the last audio chunk
• A TranscriptStabilizer compares successive transcriptions and commits stable text prefixes, preventing text flickering

Electron-Managed Runtime

Used by whisper.cpp. DeLive manages the whisper-server binary lifecycle:

1. Import or download the binary and model
2. DeLive spawns the process and waits for HTTP readiness (up to 20 seconds)
3. Audio is sent to POST /inference as WAV
4. Process is stopped on disconnect or app quit

Soniox V4

The most feature-rich provider with real-time streaming, translation, and speaker diarization.

Required: apiKey

Optional: model, languageHints, translationEnabled, translationTargetLanguage, enableSpeakerDiarization

Features:

• Token-level real-time transcription
• Real-time translation with dual-line captions
• Speaker diarization with labeled tokens
• Audio format: auto (WebM/Opus from MediaRecorder)

Volcengine (火山引擎)

Chinese-focused real-time streaming through an embedded proxy.

Required: appKey, accessKey

Optional: languageHints

The browser cannot set custom WebSocket headers, so DeLive runs an embedded HTTP proxy in the Electron main process that forwards PCM16 audio to ByteDance's openspeech.bytedance.com endpoint with the required authentication headers.

Groq

Whisper large-v3-turbo / large-v3 through Groq's high-performance inference API.

Required: apiKey

Optional: model, languageHints

SiliconFlow (硅基流动)

SenseVoice, TeleSpeech, and Qwen Omni models through SiliconFlow's API.

Required: apiKey

Optional: model, languageHints

Local OpenAI-Compatible

Works with Ollama or any service exposing the OpenAI-compatible audio transcription endpoint.

Required: baseUrl, model

Optional: apiKey, languageHints

DeLive can probe the service at baseUrl, list installed models via /v1/models, and pull models from Ollama if detected.

Local whisper.cpp

Fully offline transcription using the whisper-server binary.

Required: modelPath

Optional: binaryPath, port (default 8177), languageHints

DeLive can import or download both the binary and model files. Silent audio chunks are automatically skipped to reduce unnecessary inference.