AI/teaching-feedback-assistant
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

feat: add local voice transcription pipeline

Browse Source
This commit is contained in:
shay7sev
2026-07-20 10:49:41 +08:00
parent f1e5fd8793
commit 8a88706ff9
16 changed files with 2389 additions and 28 deletions
Download Patch File Download Diff File Expand all files Collapse all files

22
.env.example
View File

@@ -7,3 +7,25 @@ DATABASE_MAX_CONNECTIONS=5
# A single development origin. Leave unset to disable CORS middleware.
# CORS_ALLOWED_ORIGIN=https://servicewechat.com
# Private server directory used to retain uploaded audio for transcription retry.
AUDIO_STORAGE_DIR=./data/audio
# Local FunASR is the default Chinese transcription provider.
ASR_PROVIDER=local
LOCAL_ASR_URL=http://127.0.0.1:10095
ASR_REQUEST_TIMEOUT_SECONDS=1800
ASR_WORKER_CONCURRENCY=1
ASR_JOB_LEASE_SECONDS=3600
# Optional Tencent Cloud fallback. Set ASR_PROVIDER=tencent and configure all values.
# TENCENT_CLOUD_ASR_APP_ID=1234567890
# TENCENT_CLOUD_ASR_SECRET_ID=your-secret-id
# TENCENT_CLOUD_ASR_SECRET_KEY=your-secret-key
TENCENT_CLOUD_ASR_ENGINE_TYPE=16k_zh
# Optional OpenAI-compatible chat completions endpoint for polished feedback summaries.
# Without it, the API uses a local extractive summary and remains functional.
# FEEDBACK_SUMMARY_API_URL=https://api.example.com/v1/chat/completions
# FEEDBACK_SUMMARY_API_KEY=your-api-key
# FEEDBACK_SUMMARY_MODEL=your-model