Speakr is a personal, self-hosted web application designed for transcribing audio recordings
Python
Highlight cloud diarization feature and fix ASR timeout fallback
Murtaza Nasird45f48e9cd6a
diff --git a/README.md b/README.mdindex ff04f0fc..ec3f4d15 100644--- a/README.md+++ b/README.md@@ -158,14 +158,12 @@ Complete documentation is available at **[murtaza-nasir.github.io/speakr](https:## Latest Release (v0.8.0)-**Connector Architecture & REST API**+**Cloud Diarization & REST API**-- **Connector-Based Transcription** - New modular architecture for transcription providers; custom connectors possible (adding new providers requires building from source)-- **Simplified Configuration** - Fewer environment variables needed; auto-detects connector from settings-- **OpenAI Diarization Support** - Use `gpt-4o-transcribe-diarize` for speaker identification without self-hosting+- **Speaker Diarization Without GPU** - Use OpenAI's `gpt-4o-transcribe-diarize` for speaker identification with just an API key. No self-hosted ASR container or GPU required.- **REST API v1** - Full-featured API for automation tools (n8n, Zapier, Make) and dashboard widgets- **Interactive API Docs** - Swagger UI at `/api/v1/docs` with try-it-out functionality-- **Batch Operations** - Update, delete, or transcribe multiple recordings via API+- **Connector Architecture** - Modular transcription providers with simplified configuration; custom connectors possible (requires building from source)> **Migration Note:** Existing configurations are fully backwards compatible. See the [Migration Guide](https://murtaza-nasir.github.io/speakr/admin-guide/migration-guide) for recommended updates.diff --git a/src/services/transcription/registry.py b/src/services/transcription/registry.pyindex a314862f..9eec4ff4 100644--- a/src/services/transcription/registry.py+++ b/src/services/transcription/registry.py@@ -171,6 +171,34 @@ class ConnectorRegistry:logger.error(f"Failed to initialize connector '{connector_name}': {e}")raise ConfigurationError(f"Failed to initialize connector '{connector_name}': {e}") from e+ def _get_asr_timeout(self) -> int:+ """+ Get ASR timeout with fallback chain: ENV -> Admin UI -> default.++ Priority:+ 1. ASR_TIMEOUT environment variable+ 2. asr_timeout_seconds environment variable (legacy)+ 3. SystemSetting from Admin UI (database)+ 4. Default: 1800 seconds (30 minutes)+ """+ # Check environment variables first+ env_timeout = os.environ.get('ASR_TIMEOUT') or os.environ.get('asr_timeout_seconds')+ if env_timeout:+ return int(env_timeout)++ # Fall back to Admin UI setting (SystemSetting in database)+ try:+ from src.models import SystemSetting+ db_timeout = SystemSetting.get_setting('asr_timeout_seconds', None)+ if db_timeout is not None:+ return int(db_timeout)+ except Exception as e:+ # May fail if no app context or during initialization+ logger.debug(f"Could not read ASR timeout from database: {e}")++ # Default: 30 minutes+ return 1800+def _build_config_from_env(self, connector_name: str) -> Dict[str, Any]:"""Build connector config from environment variables.@@ -188,7 +216,7 @@ class ConnectorRegistry:return {'base_url': base_url,- 'timeout': int(os.environ.get('ASR_TIMEOUT', os.environ.get('asr_timeout_seconds', '1800'))),+ 'timeout': self._get_asr_timeout(),'diarize': os.environ.get('ASR_DIARIZE', 'true').lower() == 'true','return_speaker_embeddings': os.environ.get('ASR_RETURN_SPEAKER_EMBEDDINGS', 'false').lower() == 'true'}