VoiceLinkVR-server

This is a Docker server program that supports faster-whisper and WhisperLive-TRT backends

It can be used together with other clients of VoiceLinkVR, such as VRCLS, or independently as a server

Integrated a very simple user control interface. The project relies on faster-whisper-server (now called speaches) and supports WhisperLive-TRT REST integration.

Deployment

This repository uses a single compose.yaml entrypoint.

All image names, ports, cache paths, and runtime options are configured through .env, and optional services are selected through COMPOSE_PROFILES.

Quick start

Copy .env.example to .env
Update the published image names you want to use
Set COMPOSE_PROFILES to match the services you want to start
Keep WHISPER_BACKEND aligned with the enabled backend profile

cp .env.example .env
docker compose pull
docker compose up -d

Common profile combinations

COMPOSE_PROFILES=redis,faster-whisper
COMPOSE_PROFILES=redis,whisper-trt,gemma
COMPOSE_PROFILES=redis,whisper-trt,gemma,sensevoice

Important environment variables

VOICELINK_IMAGE, FASTER_WHISPER_IMAGE, WHISPER_TRT_IMAGE, GEMMA_TRANSLATE_IMAGE, SENSEVOICE_IMAGE
WHISPER_BACKEND=faster-whisper or WHISPER_BACKEND=whisper-trt
TRANSLATE_SERVICE_URL=http://gemma-translate:8080
GEMMA_TRANSLATE_ENABLED=True when enabling the Gemma service path
GEMMA_TRANSLATE_API_MODE=completion is required for the current TranslateGemma deployment (llama-server --no-jinja, POST /completion)
sourceLanguage=auto is no longer supported for Gemma/text translation requests
WHISPER_TRT_REST_HOST_PORT=18100 by default to avoid conflicting with Faster-Whisper on 8000

After everything is ready to run, please visit: http://{Server IP}:8980/ui/login

The username and password entered during the first login will be used as the default administrator's account and password. Please keep them safe.

If you forget your administrator account and password, please modify the database file yourself. In Docker, this file can be found in /usr/src/app/data/db/users.db

Service Interface Document

Management Interface

Management interface entrance: /ui/login
Management interface homepage: /ui/manage_users
Exit the management interface: /ui/logout
Delete user in the management interface: /ui/deleteUser

Control API

Except for the administrator registration interface, which does not require a token when there are no users after startup, tokens need to be added for all other interfaces

Register Administrator Interface

This interface will automatically designate the first user as the administrator when there is no user information, and the token will be verified at other times

Method: POST URL: /manageapi/registerAdmin Parameters:

{
    "username":"",
    "password":""
}

Response:

{"message": "AdminUser created successfully"}

Change User Password Interface

Method: POST URL: /manageapi/changePassword Parameters:

{
    "username":"",
    "password":""
}

Response:

{"message": "user:{username}, Password changed successfully"}

Register User Interface (Simple)

Method: POST URL: /manageapi/register Parameters:

{
    "username":"",
    "password":""
}

Response:

{"message": "User created successfully"}

Add User Interface (Full)

Method: POST URL: /manageapi/addUser Parameters:

{
    "username": "",
    "password": "",
    "is_admin": false,
    "is_active": true,
    "limit_rule": "10000/day;1000/hour",
    "expiration_date": "2025-12-31T23:59:59"
}

Field Description:

username: Username (required, 3-50 characters)
password: Password (required, minimum 6 characters)
is_admin: Is administrator (optional, default false)
is_active: Is active (optional, default true)
limit_rule: Rate limit rule (optional)
expiration_date: Expiration date (optional, ISO format)

Response:

{"message": "User created successfully", "user": {"username": ""}}

Update User Interface

Method: POST URL: /manageapi/updateUser Parameters:

{
    "username": "",
    "password": "",
    "is_admin": false,
    "is_active": true,
    "limit_rule": "10000/day;1000/hour",
    "expiration_date": "2025-12-31T23:59:59"
}

Field Description:

username: Username (required, used to locate the user to update)
password: New password (optional, not modified if not provided)
is_admin: Is administrator (optional, not modified if not provided)
is_active: Is active (optional, not modified if not provided)
limit_rule: Rate limit rule (optional, not modified if not provided)
expiration_date: Expiration date (optional, not modified if not provided)

Response:

{"message": "user:{username}, updated successfully"}

Delete User Interface

Method: POST URL: /manageapi/deleteUser Parameters:

{
    "username":""
}

Response:

{"message": "User deleted successfully"}

Call API

Login to Obtain Token

Method: POST URL: /api/login

{
    "username":"",
    "password":""
}

Response:

// Success
{"message": "Login successful", "access_token": "", "token_type": "bearer"}
// Failed
{"detail": "Invalid credentials"}

Get Speech Recognition Results (Chinese)

Method: POST URL: /api/whisper/transcriptions Parameters:

// form-data format
'file': {binary file}

Response:

{"text": ""}

Text Translation Interface

Removed in FastAPI mode. Use /api/func/webtranslate for text translation.

Request body fields: text, sourceLanguage, targetLanguage, targetLanguage2, targetLanguage3. sourceLanguage=auto is no longer supported.

Voice Translation Interface (English)

Method: POST URL: /api/func/translateToEnglish Parameters:

// form-data format
'file': {binary file}

Response:

{"text": "","translatedText":""}

Voice Translation Interface (Other Languages)

Method: POST URL: /api/func/translateToOtherLanguage Parameters:

// form-data format
files["file"]: {binary file}
data={"targetLanguage":""}
// Please refer to this project's supported language list in `src/core/services.py`
// `targetLanguage=auto` is not supported

Response:

{"text": "","translatedText":""}

Environment Variables

Variable	Description	Default
WHISPER_HOST	Whisper service address	localhost
WHISPER_PORT	Whisper service port	8000
WHISPER_BACKEND	Whisper backend selector (`faster-whisper` / `whisper-trt`)	faster-whisper
WHISPER_TRT_HOST	WhisperLive-TRT REST host	localhost
WHISPER_TRT_PORT	WhisperLive-TRT REST port	8000
SENSEVOICE_HOST	SenseVoice service address	localhost
SENSEVOICE_PORT	SenseVoice service port	8800
LIMIT_ENABLE	Enable rate limiting	False
LIMITER_REDIS_URL	Redis connection URL	-
TRANSLATOR_SERVICES_LIST	Translation services list	bing,modernMt,cloudTranslation
UPDATE_PUBLIC_BASE_URL	Public base URL for update files (optional)	-
UPDATE_STATIC_ROOT	Static root folder for update assets	src/data/update/files
UPDATE_MANIFEST_PATH	App update manifest path	src/data/update/update_manifest.json
MODEL_MANIFEST_PATH	Model catalog manifest path	src/data/update/models_manifest.json
TRANSLATION_PROFILE_MANIFEST_PATH	Translation profile manifest path	src/data/update/translation_profile_manifest.json
TRANSLATOR_RUNTIME_MANIFEST_PATH	Translator runtime manifest path	src/data/update/translator_runtime_manifest.json
TRANSLATION_CAPABILITIES_CACHE_SECONDS	Translator capability cache TTL (seconds)	1800
TRANSLATION_CAPABILITY_TIMEOUT	Capability fetch timeout per engine (seconds)	4.0
ENABLE_WEB_TRANSLATORS	Enable online translation services	True
JWT_SECRET_KEY	JWT secret key (change in production)	voicelinkvr-secret-key
SQL_PATH	Database connection string	sqlite:///data/db/users.db

Complete Environment Variable Reference

The table above is a quick summary. The following section is the complete reference for the current FastAPI server, run.py, .env.example, and compose.yaml.

Resolution order

Application runtime variables are loaded by src/core/config.py from process environment and .env
Explicit environment variables override values in .env
Values in .env override code defaults in src/core/config.py
Docker Compose also reads .env for variable interpolation in compose.yaml

FastAPI application variables

Speech recognition and translation

Variable	Description	Default
`WHISPER_HOST`	Faster-Whisper/OpenAI-compatible Whisper service host	`127.0.0.1`
`WHISPER_PORT`	Faster-Whisper/OpenAI-compatible Whisper service port	`8000`
`WHISPER_BACKEND`	Speech backend selector, supports `faster-whisper` and `whisper-trt`	`faster-whisper`
`WHISPER_TRT_HOST`	WhisperLive-TRT REST service host	`127.0.0.1`
`WHISPER_TRT_PORT`	WhisperLive-TRT REST service port	`8000`
`SENSEVOICE_HOST`	SenseVoice service host	`127.0.0.1`
`SENSEVOICE_PORT`	SenseVoice service port	`8800`
`WHISPER_APIKEY`	API key sent to compatible Whisper service when required	`something`
`WHISPER_MODEL`	Model name reported to the Whisper-compatible backend	`Systran/faster-whisper-large-v3`
`FILTER_WEB_URL`	Remote filter configuration source	`https://raw.githubusercontent.com/VoiceLinkVR/VoiceLinkServer/refs/heads/main/src/filter.json`
`ENABLE_WEB_TRANSLATORS`	Enable online translators from the `translators` library	`False`
`TRANSLATOR_SERVICE`	Preferred single translator service	`alibaba`
`TRANSLATOR_SERVICES_LIST`	Translator priority list used for fallback	`bing,iciba,alibaba,MyMemory,google`
`TRANSLATION_TIMEOUT`	Timeout for each translation request in seconds	`1.5`

Gemma translation integration

Variable	Description	Default
`GEMMA_TRANSLATE_ENABLED`	Enable TranslateGemma/OpenAI-compatible translation path	`False`
`GEMMA_TRANSLATE_URL`	TranslateGemma base host and port	`127.0.0.1:8080`
`GEMMA_TRANSLATE_API_MODE`	API mode for Gemma service (forced): `completion`	`completion`
`GEMMA_TRANSLATE_MODEL`	Model name sent to the Gemma service	`translategemma-12b-it-MP-GGUF`
`GEMMA_TRANSLATE_TIMEOUT`	Gemma request timeout in seconds	`10.0`
`GEMMA_TRANSLATE_MAX_TOKENS`	Max generated tokens for Gemma translation calls	`500`
`GEMMA_TRANSLATE_FALLBACK`	Fall back to other translators when Gemma fails	`True`

Authentication and session

Variable	Description	Default
`JWT_SECRET_KEY`	JWT signing secret; change in production	`wVLAF_13N6XL_QmP.DjkKsV`
`SESSION_SECRET_KEY`	Session middleware secret	`wVddLAF_13dsdddN6XL_QmP.DjkKsV`
`JWT_ACCESS_TOKEN_EXPIRES_MINUTES`	Access token lifetime in minutes	`10080`
`JWT_ALGORITHM`	JWT signing algorithm	`HS256`

Database, rate limiting, and compatibility

Variable	Description	Default
`SQL_PATH`	SQLAlchemy connection string	`sqlite:///{project_root}/data/db/users.db`
`SQL_POOL_SIZE`	SQLAlchemy pool size	`50`
`SQL_MAX_OVERFLOW`	SQLAlchemy max overflow connections	`150`
`SQL_POOL_RECYCLE`	SQLAlchemy recycle interval in seconds	`3600`
`LIMIT_ENABLE`	Enable request rate limiting	`True`
`LEGACY_CLIENT_COMPAT`	Enable old client compatibility behavior	`False`
`LIMIT_PUBLIC_TEST_USER`	Public test account exempt/special handling	empty
`LIMITER_REDIS_URL`	Redis URL for rate limiting storage	`redis://localhost:6379/0`

Updates, TTS, logging, and text processing

Variable	Description	Default
`TTS_URL`	External TTS service URL	empty
`TTS_TOKEN`	External TTS service token	empty
`LATEST_VERSION`	Version string returned to clients for update checks	empty
`PACKAGE_BASE_URL`	Base download URL for update packages	empty
`PACKAGE_TYPE`	Package suffix such as `.exe`	empty
`UPDATE_PUBLIC_BASE_URL`	Public base URL used when generating update links	empty
`UPDATE_STATIC_ROOT`	Local directory holding update files	`{project_root}/data/update/files`
`UPDATE_MANIFEST_PATH`	App update manifest path	`{project_root}/data/update/update_manifest.json`
`MODEL_MANIFEST_PATH`	Model manifest path	`{project_root}/data/update/models_manifest.json`
`TRANSLATION_PROFILE_MANIFEST_PATH`	Translation profile manifest path	`{project_root}/data/update/translation_profile_manifest.json`
`TRANSLATOR_RUNTIME_MANIFEST_PATH`	Translator runtime manifest path	`{project_root}/data/update/translator_runtime_manifest.json`
`TRANSLATION_CAPABILITIES_CACHE_SECONDS`	Capability cache TTL in seconds	`1800`
`TRANSLATION_CAPABILITY_TIMEOUT`	Capability probe timeout per engine in seconds	`4.0`
`LOG_LEVEL`	Application log level	`INFO`
`ENABLE_TEXT_COMPRESSION`	Enable repeated-character text compression	`True`
`TEXT_COMPRESSION_MIN_REPEAT`	Minimum repeat count before compression	`5`

Afdian membership/payment

Variable	Description	Default
`AFDIAN_USER_ID`	Afdian user ID	`db43f694299511ef821252540025c377`
`AFDIAN_TOKEN`	Afdian API token	empty
`AFDIAN_ORDER_CREATE_BASE_URL`	Afdian payment page base URL	`https://ifdian.net/p/9c65d9cc617011ed81c352540025c377`
`AFDIAN_QUERY_ORDER_URL`	Afdian order query API	`https://afdian.com/api/open/query-order`
`AFDIAN_CUSTOM_ID_KEY`	Secret used to sign/encrypt custom IDs	`change-me-to-32-byte-secret`
`AFDIAN_CUSTOM_PRICE`	Base custom price	`5.0`
`AFDIAN_RENEW_DAYS`	Days granted for one renewal unit	`30`
`AFDIAN_BOOST_DAYS`	Days used by boost plan calculations	`30`
`AFDIAN_BOOST_PRICE_STEP`	Price step for boost upgrades	`5.0`
`AFDIAN_BOOST_DAILY_INCREMENT`	Daily quota increment for each boost step	`1500`
`AFDIAN_DEFAULT_LIMIT_RULE`	Default quota/rate rule for paid users	`1500/day`
`AFDIAN_BOOST_LIMIT_RULE`	Optional explicit quota rule for boost users	empty
`AFDIAN_QUERY_SCAN_PAGES`	Number of order pages to scan per sync	`3`
`AFDIAN_CUSTOM_ID_TTL_SECONDS`	Custom ID validity period in seconds	`604800`
`AFDIAN_WEBHOOK_FORWARD_ENABLED`	Whether to transparently forward Afdian webhooks to the bot plugin	`False`
`AFDIAN_WEBHOOK_FORWARD_URL`	Bot plugin webhook URL	empty
`AFDIAN_WEBHOOK_FORWARD_TIMEOUT`	Forward timeout in seconds	`10.0`
`AFDIAN_WEBHOOK_FORWARD_AUTHORIZATION`	Authorization header sent during forwarding	empty

Afdian webhook and Nginx reverse proxy

The recommended flow is now:

Afdian only calls the server webhook at /api/user-center/afdian/webhook
The server immediately returns {"ec":200,"em":""} to acknowledge receipt
The server then processes activation in the background and transparently forwards the original webhook JSON to the bot plugin

Example configuration:

AFDIAN_WEBHOOK_FORWARD_ENABLED=True
AFDIAN_WEBHOOK_FORWARD_URL=http://127.0.0.1:6500/
AFDIAN_WEBHOOK_FORWARD_TIMEOUT=10
AFDIAN_WEBHOOK_FORWARD_AUTHORIZATION=

Example Nginx config:

server {
    listen 443 ssl http2;
    server_name your-domain.example;

    # Only expose the server webhook endpoint to Afdian
    location = /api/user-center/afdian/webhook {
        proxy_pass http://127.0.0.1:8980/api/user-center/afdian/webhook;
        proxy_set_header Host $host;
        proxy_set_header X-Real-IP $remote_addr;
        proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
        proxy_set_header X-Forwarded-Proto $scheme;
    }
}

Set the endpoints like this:

Afdian webhook URL: https://your-domain.example/api/user-center/afdian/webhook
Server forwarding target AFDIAN_WEBHOOK_FORWARD_URL: the bot plugin's real reachable address, for example http://127.0.0.1:6500/ or http://192.168.1.10:6500/

If the bot plugin runs on the same machine or on a private network, it does not need its own public Nginx entry.

Uvicorn startup variables

Variable	Description	Default
`UVICORN_WORKERS`	Worker process count used by `run.py`	`4`
`UVICORN_RELOAD`	Enable hot reload in `run.py`; `true/1/t` means enabled	`False`

Docker Compose variables

These variables are read by compose.yaml and usually come from .env.

Compose control and naming

Variable	Description	Default
`COMPOSE_PROFILES`	Optional service profiles to start	`redis,faster-whisper`
`COMPOSE_PROJECT_NAME`	Compose project name	`voicelinkvr`
`REDIS_CONTAINER_NAME`	Redis container name override	`redis`
`FASTER_WHISPER_CONTAINER_NAME`	Faster-Whisper container name override	`faster-whisper`
`WHISPER_TRT_CONTAINER_NAME`	WhisperLive-TRT container name override	`whisper-trt`
`SENSEVOICE_CONTAINER_NAME`	SenseVoice container name override	`sensevoice`
`GEMMA_TRANSLATE_CONTAINER_NAME`	Gemma container name override	`gemma-translate`
`VOICELINK_CONTAINER_NAME`	VoiceLink app container name override	`voicelink`

Image sources

Variable	Description	Default
`VOICELINK_IMAGE`	VoiceLink server image	`boyqiu0010/voice-link-vr-server:latest`
`REDIS_IMAGE`	Redis image	`redis:latest`
`FASTER_WHISPER_IMAGE`	Faster-Whisper image	`fedirz/faster-whisper-server:latest-cuda`
`WHISPER_TRT_IMAGE`	WhisperLive-TRT image	`ghcr.io/your-org/whisperlive-trt-gemma:latest`
`GEMMA_TRANSLATE_IMAGE`	TranslateGemma image	`ghcr.io/your-org/translategemma-llama-server:latest`
`SENSEVOICE_IMAGE`	SenseVoice image	`ghcr.io/your-org/sensevoice-oneapi:latest`

Published ports

Variable	Description	Default
`VOICELINK_HOST_PORT`	Published port for the VoiceLink server	`8980`
`REDIS_HOST_PORT`	Published port for Redis	`6379`
`FASTER_WHISPER_HOST_PORT`	Published port for Faster-Whisper	`8000`
`WHISPER_TRT_WS_HOST_PORT`	Published WebSocket port for WhisperLive-TRT	`9090`
`WHISPER_TRT_REST_HOST_PORT`	Published REST port for WhisperLive-TRT	`18100`
`GEMMA_TRANSLATE_BIND_IP`	Bind address used when exposing Gemma	`127.0.0.1`
`GEMMA_TRANSLATE_HOST_PORT`	Published port for Gemma	`8080`
`SENSEVOICE_HOST_PORT`	Published port for SenseVoice	`8800`

Volumes and local paths

Variable	Description	Default
`REDIS_DATA_PATH`	Host path mounted to Redis `/data`	`./redisFiles/data`
`REDIS_CONFIG_PATH`	Host path mounted as Redis config	`./redisFiles/redis.conf`
`VOICELINK_DATA_VOLUME`	Named volume for `/usr/src/app/data`	`voice-link-vr-server-local`
`WL_CACHE_DIR`	Host cache root for WhisperLive-TRT assets/engine/weights	`./.cache/whisper-trt`
`GEMMA_MODEL_ROOT`	Host model directory mounted to `/models`	`D:/models`

Faster-Whisper container variables

Variable	Description	Default
`FASTER_WHISPER_TTL`	Model/process TTL inside Faster-Whisper container	`-1`
`FASTER_WHISPER_NUM_WORKERS`	Worker count inside Faster-Whisper container	`10`
`FASTER_WHISPER_CPU_THREADS`	CPU thread count inside Faster-Whisper container	`40`

WhisperLive-TRT container variables

Variable	Description	Default
`WHISPER_TRT_ENABLE_REST`	Enable REST API in WhisperLive-TRT	`true`
`WHISPER_TRT_REST_PORT`	Internal REST port in WhisperLive-TRT container	`8000`
`WHISPER_TRT_MULTILINGUAL`	Enable multilingual mode	`true`
`TRANSLATE_SERVICE_URL`	Translation service URL used by WhisperLive-TRT	`http://gemma-translate:8080`
`AUTO_BUILD_TRT_MODEL`	Auto-build TensorRT model on startup	`true`
`TRT_BUILD_MODEL_NAME`	Whisper model name used when building TRT engine	`large-v3`
`TRT_WEIGHT_ONLY_PRECISION`	Weight precision used during TRT build	`float16`
`TRT_PY_SESSION`	Enable Python TRT session mode	`false`
`TRT_BUILD_INFERENCE_PRECISION`	Inference precision used at build time	`float16`
`TRT_BUILD_MAX_BATCH_SIZE`	Max batch size for build	`1`
`TRT_BUILD_MAX_BEAM_WIDTH`	Max beam width for build	`1`
`TRT_BUILD_ENCODER_MAX_INPUT_LEN`	Max encoder input length during build	`3000`
`TRT_BUILD_ENCODER_MAX_SEQ_LEN`	Max encoder sequence length during build	`3000`
`TRT_BUILD_DECODER_MAX_SEQ_LEN`	Max decoder sequence length during build	`225`
`TRT_BUILD_DECODER_MAX_INPUT_LEN`	Max decoder input length during build	`32`
`TRT_BUILD_DECODER_MAX_ENCODER_INPUT_LEN`	Max decoder encoder input length during build	`3000`
`TRT_BUILD_DECODER_MAX_NUM_TOKENS`	Optional decoder token cap during build	empty
`TRT_RUNTIME_MAX_BATCH_SIZE`	Max batch size at runtime	`1`
`TRT_RUNTIME_MAX_INPUT_LEN`	Max runtime input length	`3000`
`TRT_RUNTIME_MAX_OUTPUT_LEN`	Max runtime output length	`64`
`TRT_RUNTIME_MAX_BEAM_WIDTH`	Max runtime beam width	`1`
`TRT_KV_CACHE_FREE_GPU_MEMORY_FRACTION`	Fraction of free GPU memory reserved for KV cache	`0.55`
`TRT_CROSS_KV_CACHE_FRACTION`	Fraction reserved for cross KV cache	`0.30`
`CONTEXT_FMHA_MODE`	Context FMHA plugin mode	`auto`
`BERT_ATTENTION_PLUGIN_MODE`	BERT attention plugin mode	`auto`
`GPT_ATTENTION_PLUGIN_MODE`	GPT attention plugin mode	`auto`
`CORS_ORIGINS`	Allowed CORS origins for WhisperLive-TRT REST	`http://127.0.0.1:18100,http://localhost:18100`

TranslateGemma container variables

Variable	Description	Default
`GEMMA_MODEL_NAME`	Model loaded by the Gemma container	`translategemma-12b-it-MP-GGUF`
`LLAMA_N_GPU_LAYERS`	Number of model layers kept on GPU	`-1`
`LLAMA_N_PARALLEL`	Parallel request count for llama-server	`2`
`LLAMA_N_CTX`	Context length for llama-server	`3072`
`LLAMA_FLASH_ATTN`	Enable flash attention	`on`
`LLAMA_EXTRA_ARGS`	Extra startup args passed to llama-server	empty

SenseVoice container variables

Variable	Description	Default
`SENSEVOICE_DEVICE_TYPE`	Device passed to SenseVoice	`cuda:0`
`SENSEVOICE_USE_ITN`	Enable inverse text normalization	`True`
`SENSEVOICE_NCPU`	CPU count used by SenseVoice	`8`
`SENSEVOICE_VAD_ENABLE`	Enable VAD in SenseVoice	`False`

Compatibility and legacy variables

The repository still contains older Flask/Werkzeug files (src/server.py, src/serverstart.py) that reference some historical variables. They are not part of the current FastAPI startup path (run.py -> src/main.py), but you may still encounter them in old deployments.

Variable	Status	Notes
`LIBRETRANSLATE_HOST`	legacy	Used by the old Flask server code
`LIBRETRANSLATE_PORT`	legacy	Used by the old Flask server code
`LIBRETRANSLATE_APIKEY`	legacy	Used by the old Flask server code
`FLASK_SECRET_KEY`	legacy	Replaced by `SESSION_SECRET_KEY` in FastAPI
`JWT_ACCESS_TOKEN_EXPIRES`	legacy	Replaced by `JWT_ACCESS_TOKEN_EXPIRES_MINUTES`
`SQLALCHEMY_POOL_SIZE`	legacy/compose	Historical name, current FastAPI setting is `SQL_POOL_SIZE`
`SQLALCHEMY_MAX_OVERFLOW`	legacy	Historical name, current FastAPI setting is `SQL_MAX_OVERFLOW`
`THREADS_NUM`	legacy	Used by old Waitress startup script

Afdian webhook reverse proxy correction

If you only need one public webhook entry for Afdian, Nginx should reverse proxy directly to the server webhook endpoint, not to the bot plugin.

Recommended settings:

AFDIAN_WEBHOOK_FORWARD_ENABLED=True
AFDIAN_WEBHOOK_FORWARD_URL=http://127.0.0.1:6500/
AFDIAN_WEBHOOK_FORWARD_TIMEOUT=10
AFDIAN_WEBHOOK_FORWARD_AUTHORIZATION=

server {
    listen 443 ssl http2;
    server_name your-domain.example;

    location = /api/user-center/afdian/webhook {
        proxy_pass http://127.0.0.1:8980/api/user-center/afdian/webhook;
        proxy_set_header Host $host;
        proxy_set_header X-Real-IP $remote_addr;
        proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
        proxy_set_header X-Forwarded-Proto $scheme;
    }
}

Use these values:

Afdian webhook URL: https://your-domain.example/api/user-center/afdian/webhook
AFDIAN_WEBHOOK_FORWARD_URL: the bot plugin's real reachable address, for example http://127.0.0.1:6500/ or a private LAN address

If the bot plugin runs on the same host or inside a private network, it does not need its own public Nginx endpoint.

Name		Name	Last commit message	Last commit date
Latest commit History 70 Commits
.playwright-cli		.playwright-cli
docs		docs
redisFiles		redisFiles
src		src
tmp		tmp
tools		tools
.dockerignore		.dockerignore
.env.example		.env.example
.gitignore		.gitignore
AGENTS.md		AGENTS.md
CLAUDE.md		CLAUDE.md
Dockerfile		Dockerfile
LICENSE		LICENSE
README-zh.md		README-zh.md
README.md		README.md
README_FastAPI.md		README_FastAPI.md
README_TEXT_COMPRESSION.md		README_TEXT_COMPRESSION.md
WhisperLive-TRT集成方案.md		WhisperLive-TRT集成方案.md
check_db.py		check_db.py
compose.yaml		compose.yaml
run.py		run.py
test_translate.py		test_translate.py
修改文档.md		修改文档.md
性能提高.md		性能提高.md
服务器请求计数调整预告.md		服务器请求计数调整预告.md
迁移文档.md		迁移文档.md

Folders and files

Latest commit

History

Repository files navigation

VoiceLinkVR-server

Deployment

Quick start

Common profile combinations

Important environment variables

Service Interface Document

Management Interface

Control API

Register Administrator Interface

Change User Password Interface

Register User Interface (Simple)

Add User Interface (Full)

Update User Interface

Delete User Interface

Call API

Login to Obtain Token

Get Speech Recognition Results (Chinese)

Text Translation Interface

Voice Translation Interface (English)

Voice Translation Interface (Other Languages)

Environment Variables

Complete Environment Variable Reference

Resolution order

FastAPI application variables

Speech recognition and translation

Gemma translation integration

Authentication and session

Database, rate limiting, and compatibility

Updates, TTS, logging, and text processing

Afdian membership/payment

Afdian webhook and Nginx reverse proxy

Uvicorn startup variables

Docker Compose variables

Compose control and naming

Image sources

Published ports

Volumes and local paths

Faster-Whisper container variables

WhisperLive-TRT container variables

TranslateGemma container variables

SenseVoice container variables

Compatibility and legacy variables

Afdian webhook reverse proxy correction

About

Topics

Resources

License

Uh oh!

Stars

Watchers

Forks

Releases

Packages 0

Uh oh!

Uh oh!

Contributors

Uh oh!

Languages

Packages