An error is displayed in the Open WebUI

[Hunter6324]

It can be tested by a python program, but in the open webui it prompts:
[ERROR: 400: [ERROR: 'text']]

[gitlawr]

Can you provide OpenWebUI configuration screenshot or steps to help us check?

[Hunter6324]

terminal display：

PS C:\Users\tom\Python_Project\vox-box> C:\Users\tom\Python_Project\vox-box\start_vox_box.bat
2025-06-12T16:37:33+08:00 - vox_box.cmd.start - INFO - Starting with arguments: [('debug', None), ('host', '0.0.0.0'), ('port', 8080), ('model', None), ('device', 'cpu'), ('huggingface_repo_id', None), ('model_scope_model_id', 'iic/speech_paraformer-large-vad-punc_asr_nat-zh-cn-16k-common-vocab8404-pytorch'), ('data_dir', 'C:\Users\tom\Python_Project\vox-box\downloads'), ('func', <function run at 0x00000228D3CC6160>)]
2025-06-12T16:37:33+08:00 - vox_box.server.model - INFO - Estimating model
Downloading Model to directory: C:\Users\tom\Python_Project\vox-box\downloads\cache\model_scope\hub\iic/speech_paraformer-large-vad-punc_asr_nat-zh-cn-16k-common-vocab8404-pytorch
2025-06-12T16:37:34+08:00 - modelscope - WARNING - Model revision not specified, use revision: v2.0.5
Downloading Model to directory: C:\Users\tom\Python_Project\vox-box\downloads\cache\model_scope\hub\iic/speech_paraformer-large-vad-punc_asr_nat-zh-cn-16k-common-vocab8404-pytorch
2025-06-12T16:37:36+08:00 - modelscope - WARNING - Model revision not specified, use revision: v2.0.5
2025-06-12T16:37:36+08:00 - vox_box.server.model - INFO - Finished estimating model
2025-06-12T16:37:36+08:00 - vox_box.server.model - INFO - Downloading model
Downloading Model to directory: C:\Users\tom\Python_Project\vox-box\downloads\cache\model_scope\hub\iic/speech_paraformer-large-vad-punc_asr_nat-zh-cn-16k-common-vocab8404-pytorch
2025-06-12T16:37:36+08:00 - modelscope - WARNING - Model revision not specified, use revision: v2.0.5
2025-06-12T16:37:37+08:00 - vox_box.server.model - INFO - Loading model
2025-06-12T16:37:38+08:00 - root - INFO - New registry table added: preprocessor_classes
Key Conformer already exists in model_classes, re-register
2025-06-12T16:37:41+08:00 - root - INFO - New registry table added: adaptor_classes
Key Linear already exists in adaptor_classes, re-register
Key TransformerDecoder already exists in decoder_classes, re-register
Key LightweightConvolutionTransformerDecoder already exists in decoder_classes, re-register
Key LightweightConvolution2DTransformerDecoder already exists in decoder_classes, re-register
Key DynamicConvolutionTransformerDecoder already exists in decoder_classes, re-register
Key DynamicConvolution2DTransformerDecoder already exists in decoder_classes, re-register
2025-06-12T16:37:43+08:00 - root - INFO - New registry table added: lid_predictor_classes
funasr version: 1.1.14.
2025-06-12T16:37:44+08:00 - root - INFO - download models from model hub: ms
2025-06-12T16:37:47+08:00 - root - INFO - Loading pretrained params from C:\Users\tom\Python_Project\vox-box\downloads\cache\model_scope\iic\speech_paraformer-large-vad-punc_asr_nat-zh-cn-16k-common-vocab8404-pytorch\model.pt
2025-06-12T16:37:47+08:00 - root - INFO - ckpt: C:\Users\tom\Python_Project\vox-box\downloads\cache\model_scope\iic\speech_paraformer-large-vad-punc_asr_nat-zh-cn-16k-common-vocab8404-pytorch\model.pt
2025-06-12T16:37:48+08:00 - root - INFO - scope_map: ['module.', 'None']
2025-06-12T16:37:48+08:00 - root - INFO - excludes: None
2025-06-12T16:37:49+08:00 - root - INFO - Loading ckpt: C:\Users\tom\Python_Project\vox-box\downloads\cache\model_scope\iic\speech_paraformer-large-vad-punc_asr_nat-zh-cn-16k-common-vocab8404-pytorch\model.pt, status:
2025-06-12T16:37:49+08:00 - vox_box.server.server - INFO - Starting Vox Box server.
2025-06-12T16:37:49+08:00 - vox_box.server.server - INFO - Serving on 0.0.0.0:8080.
2025-06-12T16:38:53+08:00 - python_multipart.multipart - WARNING - Skipping data after last boundary

[Hunter6324]

More detailed terminal display

PS C:\Users\tom\Python_Project\vox-box> vox-box start --model-scope-model-id iic/speech_paraformer-large-vad-punc_asr_nat-zh-cn-16k-common-vocab8404-pytorch --data-dir C:\Users\tom\Python_Project\vox-box\downloads --host 0.0.0.0 --port 8080 --device cpu -d
2025-06-12T16:44:34+08:00 - vox_box.cmd.start - INFO - Starting with arguments: [('debug', True), ('host', '0.0.0.0'), ('port', 8080), ('model', None), ('device', 'cpu'), ('huggingface_repo_id', None), ('model_scope_model_id', 'iic/speech_paraformer-large-vad-punc_asr_nat-zh-cn-16k-common-vocab8404-pytorch'), ('data_dir', 'C:\Users\tom\Python_Project\vox-box\downloads'), ('func', <function run at 0x0000021DE32C6160>)]
2025-06-12T16:44:34+08:00 - vox_box.server.model - INFO - Estimating model
2025-06-12T16:44:34+08:00 - vox_box.downloader.downloaders - DEBUG - Downloading modelscope:iic/speech_paraformer-large-vad-punc_asr_nat-zh-cn-16k-common-vocab8404-pytorch:cosyvoice.yaml
2025-06-12T16:44:34+08:00 - filelock - DEBUG - Attempting to acquire lock 2327388771664 on C:\Users\tom\Python_Project\vox-box\downloads\cache\model_scope\iic\speech_paraformer-large-vad-punc_asr_nat-zh-cn-16k-common-vocab8404-pytorch.lock
2025-06-12T16:44:34+08:00 - filelock - DEBUG - Lock 2327388771664 acquired on C:\Users\tom\Python_Project\vox-box\downloads\cache\model_scope\iic\speech_paraformer-large-vad-punc_asr_nat-zh-cn-16k-common-vocab8404-pytorch.lock
2025-06-12T16:44:34+08:00 - urllib3.connectionpool - DEBUG - Starting new HTTPS connection (1): www.modelscope.cn:443
2025-06-12T16:44:34+08:00 - urllib3.connectionpool - DEBUG - https://www.modelscope.cn:443 "GET /api/v1/models/iic/speech_paraformer-large-vad-punc_asr_nat-zh-cn-16k-common-vocab8404-pytorch/repo/files?Recursive=False HTTP/11" 200 None
2025-06-12T16:44:34+08:00 - filelock - DEBUG - Attempting to release lock 2327388771664 on C:\Users\tom\Python_Project\vox-box\downloads\cache\model_scope\iic\speech_paraformer-large-vad-punc_asr_nat-zh-cn-16k-common-vocab8404-pytorch.lock
2025-06-12T16:44:34+08:00 - filelock - DEBUG - Lock 2327388771664 released on C:\Users\tom\Python_Project\vox-box\downloads\cache\model_scope\iic\speech_paraformer-large-vad-punc_asr_nat-zh-cn-16k-common-vocab8404-pytorch.lock
2025-06-12T16:44:34+08:00 - vox_box.estimator.cosyvoice - DEBUG - File cosyvoice.yaml does not exist, No file found in iic/speech_paraformer-large-vad-punc_asr_nat-zh-cn-16k-common-vocab8404-pytorch that match cosyvoice.yaml
2025-06-12T16:44:35+08:00 - urllib3.connectionpool - DEBUG - Starting new HTTPS connection (1): www.modelscope.cn:443
2025-06-12T16:44:35+08:00 - urllib3.connectionpool - DEBUG - https://www.modelscope.cn:443 "GET /api/v1/models/iic/speech_paraformer-large-vad-punc_asr_nat-zh-cn-16k-common-vocab8404-pytorch/repo/files?Recursive=False HTTP/11" 200 None
2025-06-12T16:44:35+08:00 - vox_box.downloader.downloaders - DEBUG - Downloading modelscope:iic/speech_paraformer-large-vad-punc_asr_nat-zh-cn-16k-common-vocab8404-pytorch:configuration.json
2025-06-12T16:44:35+08:00 - filelock - DEBUG - Attempting to acquire lock 2327388769232 on C:\Users\tom\Python_Project\vox-box\downloads\cache\model_scope\iic\speech_paraformer-large-vad-punc_asr_nat-zh-cn-16k-common-vocab8404-pytorch.lock
2025-06-12T16:44:35+08:00 - filelock - DEBUG - Lock 2327388769232 acquired on C:\Users\tom\Python_Project\vox-box\downloads\cache\model_scope\iic\speech_paraformer-large-vad-punc_asr_nat-zh-cn-16k-common-vocab8404-pytorch.lock
2025-06-12T16:44:35+08:00 - urllib3.connectionpool - DEBUG - Starting new HTTPS connection (1): www.modelscope.cn:443
2025-06-12T16:44:35+08:00 - urllib3.connectionpool - DEBUG - https://www.modelscope.cn:443 "GET /api/v1/models/iic/speech_paraformer-large-vad-punc_asr_nat-zh-cn-16k-common-vocab8404-pytorch/repo/files?Recursive=False HTTP/11" 200 None
Downloading Model to directory: C:\Users\tom\Python_Project\vox-box\downloads\cache\model_scope\hub\iic/speech_paraformer-large-vad-punc_asr_nat-zh-cn-16k-common-vocab8404-pytorch
2025-06-12T16:44:35+08:00 - urllib3.connectionpool - DEBUG - Starting new HTTPS connection (1): www.modelscope.cn:443
2025-06-12T16:44:36+08:00 - urllib3.connectionpool - DEBUG - https://www.modelscope.cn:443 "GET /api/v1/models/iic/speech_paraformer-large-vad-punc_asr_nat-zh-cn-16k-common-vocab8404-pytorch/revisions HTTP/11" 200 1267
2025-06-12T16:44:36+08:00 - modelscope - WARNING - Model revision not specified, use revision: v2.0.5
2025-06-12T16:44:36+08:00 - urllib3.connectionpool - DEBUG - https://www.modelscope.cn:443 "GET /api/v1/models/iic/speech_paraformer-large-vad-punc_asr_nat-zh-cn-16k-common-vocab8404-pytorch/repo/files?Revision=v2.0.5&Recursive=True HTTP/11" 200 None
2025-06-12T16:44:36+08:00 - filelock - DEBUG - Attempting to release lock 2327388769232 on C:\Users\tom\Python_Project\vox-box\downloads\cache\model_scope\iic\speech_paraformer-large-vad-punc_asr_nat-zh-cn-16k-common-vocab8404-pytorch.lock
2025-06-12T16:44:36+08:00 - filelock - DEBUG - Lock 2327388769232 released on C:\Users\tom\Python_Project\vox-box\downloads\cache\model_scope\iic\speech_paraformer-large-vad-punc_asr_nat-zh-cn-16k-common-vocab8404-pytorch.lock
2025-06-12T16:44:36+08:00 - vox_box.downloader.downloaders - DEBUG - Downloaded modelscope:iic/speech_paraformer-large-vad-punc_asr_nat-zh-cn-16k-common-vocab8404-pytorch:configuration.json
2025-06-12T16:44:36+08:00 - vox_box.downloader.downloaders - DEBUG - Downloading modelscope:iic/speech_paraformer-large-vad-punc_asr_nat-zh-cn-16k-common-vocab8404-pytorch:config.json
2025-06-12T16:44:36+08:00 - filelock - DEBUG - Attempting to acquire lock 2327388830736 on C:\Users\tom\Python_Project\vox-box\downloads\cache\model_scope\iic\speech_paraformer-large-vad-punc_asr_nat-zh-cn-16k-common-vocab8404-pytorch.lock
2025-06-12T16:44:36+08:00 - filelock - DEBUG - Lock 2327388830736 acquired on C:\Users\tom\Python_Project\vox-box\downloads\cache\model_scope\iic\speech_paraformer-large-vad-punc_asr_nat-zh-cn-16k-common-vocab8404-pytorch.lock
2025-06-12T16:44:36+08:00 - urllib3.connectionpool - DEBUG - Starting new HTTPS connection (1): www.modelscope.cn:443
2025-06-12T16:44:37+08:00 - urllib3.connectionpool - DEBUG - https://www.modelscope.cn:443 "GET /api/v1/models/iic/speech_paraformer-large-vad-punc_asr_nat-zh-cn-16k-common-vocab8404-pytorch/repo/files?Recursive=False HTTP/11" 200 None
2025-06-12T16:44:37+08:00 - filelock - DEBUG - Attempting to release lock 2327388830736 on C:\Users\tom\Python_Project\vox-box\downloads\cache\model_scope\iic\speech_paraformer-large-vad-punc_asr_nat-zh-cn-16k-common-vocab8404-pytorch.lock
2025-06-12T16:44:37+08:00 - filelock - DEBUG - Lock 2327388830736 released on C:\Users\tom\Python_Project\vox-box\downloads\cache\model_scope\iic\speech_paraformer-large-vad-punc_asr_nat-zh-cn-16k-common-vocab8404-pytorch.lock
2025-06-12T16:44:37+08:00 - vox_box.estimator.funasr - DEBUG - File config.json does not exist, No file found in iic/speech_paraformer-large-vad-punc_asr_nat-zh-cn-16k-common-vocab8404-pytorch that match config.json
2025-06-12T16:44:37+08:00 - vox_box.downloader.downloaders - DEBUG - Downloading modelscope:iic/speech_paraformer-large-vad-punc_asr_nat-zh-cn-16k-common-vocab8404-pytorch:config.yaml
2025-06-12T16:44:37+08:00 - filelock - DEBUG - Attempting to acquire lock 2327388877456 on C:\Users\tom\Python_Project\vox-box\downloads\cache\model_scope\iic\speech_paraformer-large-vad-punc_asr_nat-zh-cn-16k-common-vocab8404-pytorch.lock
2025-06-12T16:44:37+08:00 - filelock - DEBUG - Lock 2327388877456 acquired on C:\Users\tom\Python_Project\vox-box\downloads\cache\model_scope\iic\speech_paraformer-large-vad-punc_asr_nat-zh-cn-16k-common-vocab8404-pytorch.lock
2025-06-12T16:44:37+08:00 - urllib3.connectionpool - DEBUG - Starting new HTTPS connection (1): www.modelscope.cn:443
2025-06-12T16:44:37+08:00 - urllib3.connectionpool - DEBUG - https://www.modelscope.cn:443 "GET /api/v1/models/iic/speech_paraformer-large-vad-punc_asr_nat-zh-cn-16k-common-vocab8404-pytorch/repo/files?Recursive=False HTTP/11" 200 None
Downloading Model to directory: C:\Users\tom\Python_Project\vox-box\downloads\cache\model_scope\hub\iic/speech_paraformer-large-vad-punc_asr_nat-zh-cn-16k-common-vocab8404-pytorch
2025-06-12T16:44:37+08:00 - urllib3.connectionpool - DEBUG - Starting new HTTPS connection (1): www.modelscope.cn:443
2025-06-12T16:44:37+08:00 - urllib3.connectionpool - DEBUG - https://www.modelscope.cn:443 "GET /api/v1/models/iic/speech_paraformer-large-vad-punc_asr_nat-zh-cn-16k-common-vocab8404-pytorch/revisions HTTP/11" 200 1267
2025-06-12T16:44:37+08:00 - modelscope - WARNING - Model revision not specified, use revision: v2.0.5
2025-06-12T16:44:37+08:00 - urllib3.connectionpool - DEBUG - https://www.modelscope.cn:443 "GET /api/v1/models/iic/speech_paraformer-large-vad-punc_asr_nat-zh-cn-16k-common-vocab8404-pytorch/repo/files?Revision=v2.0.5&Recursive=True HTTP/11" 200 None
2025-06-12T16:44:37+08:00 - filelock - DEBUG - Attempting to release lock 2327388877456 on C:\Users\tom\Python_Project\vox-box\downloads\cache\model_scope\iic\speech_paraformer-large-vad-punc_asr_nat-zh-cn-16k-common-vocab8404-pytorch.lock
2025-06-12T16:44:37+08:00 - filelock - DEBUG - Lock 2327388877456 released on C:\Users\tom\Python_Project\vox-box\downloads\cache\model_scope\iic\speech_paraformer-large-vad-punc_asr_nat-zh-cn-16k-common-vocab8404-pytorch.lock
2025-06-12T16:44:37+08:00 - vox_box.downloader.downloaders - DEBUG - Downloaded modelscope:iic/speech_paraformer-large-vad-punc_asr_nat-zh-cn-16k-common-vocab8404-pytorch:config.yaml
2025-06-12T16:44:37+08:00 - vox_box.server.model - INFO - Finished estimating model
2025-06-12T16:44:37+08:00 - vox_box.server.model - INFO - Downloading model
2025-06-12T16:44:37+08:00 - vox_box.downloader.downloaders - DEBUG - Downloading modelscope:iic/speech_paraformer-large-vad-punc_asr_nat-zh-cn-16k-common-vocab8404-pytorch
2025-06-12T16:44:37+08:00 - filelock - DEBUG - Attempting to acquire lock 2327388685136 on C:\Users\tom\Python_Project\vox-box\downloads\cache\model_scope\iic\speech_paraformer-large-vad-punc_asr_nat-zh-cn-16k-common-vocab8404-pytorch.lock
2025-06-12T16:44:37+08:00 - filelock - DEBUG - Lock 2327388685136 acquired on C:\Users\tom\Python_Project\vox-box\downloads\cache\model_scope\iic\speech_paraformer-large-vad-punc_asr_nat-zh-cn-16k-common-vocab8404-pytorch.lock
Downloading Model to directory: C:\Users\tom\Python_Project\vox-box\downloads\cache\model_scope\hub\iic/speech_paraformer-large-vad-punc_asr_nat-zh-cn-16k-common-vocab8404-pytorch
2025-06-12T16:44:37+08:00 - urllib3.connectionpool - DEBUG - Starting new HTTPS connection (1): www.modelscope.cn:443
2025-06-12T16:44:38+08:00 - urllib3.connectionpool - DEBUG - https://www.modelscope.cn:443 "GET /api/v1/models/iic/speech_paraformer-large-vad-punc_asr_nat-zh-cn-16k-common-vocab8404-pytorch/revisions HTTP/11" 200 1267
2025-06-12T16:44:38+08:00 - modelscope - WARNING - Model revision not specified, use revision: v2.0.5
2025-06-12T16:44:38+08:00 - urllib3.connectionpool - DEBUG - https://www.modelscope.cn:443 "GET /api/v1/models/iic/speech_paraformer-large-vad-punc_asr_nat-zh-cn-16k-common-vocab8404-pytorch/repo/files?Revision=v2.0.5&Recursive=True HTTP/11" 200 None
2025-06-12T16:44:38+08:00 - filelock - DEBUG - Attempting to release lock 2327388685136 on C:\Users\tom\Python_Project\vox-box\downloads\cache\model_scope\iic\speech_paraformer-large-vad-punc_asr_nat-zh-cn-16k-common-vocab8404-pytorch.lock
2025-06-12T16:44:38+08:00 - filelock - DEBUG - Lock 2327388685136 released on C:\Users\tom\Python_Project\vox-box\downloads\cache\model_scope\iic\speech_paraformer-large-vad-punc_asr_nat-zh-cn-16k-common-vocab8404-pytorch.lock
2025-06-12T16:44:38+08:00 - vox_box.downloader.downloaders - DEBUG - Downloaded modelscope:iic/speech_paraformer-large-vad-punc_asr_nat-zh-cn-16k-common-vocab8404-pytorch
2025-06-12T16:44:38+08:00 - vox_box.server.model - INFO - Loading model
2025-06-12T16:44:39+08:00 - root - INFO - New registry table added: preprocessor_classes
Key Conformer already exists in model_classes, re-register
2025-06-12T16:44:39+08:00 - root - INFO - New registry table added: adaptor_classes
Key Linear already exists in adaptor_classes, re-register
Key TransformerDecoder already exists in decoder_classes, re-register
Key LightweightConvolutionTransformerDecoder already exists in decoder_classes, re-register
Key LightweightConvolution2DTransformerDecoder already exists in decoder_classes, re-register
Key DynamicConvolutionTransformerDecoder already exists in decoder_classes, re-register
Key DynamicConvolution2DTransformerDecoder already exists in decoder_classes, re-register
2025-06-12T16:44:40+08:00 - root - INFO - New registry table added: lid_predictor_classes
funasr version: 1.1.14.
2025-06-12T16:44:41+08:00 - root - INFO - download models from model hub: ms
2025-06-12T16:44:42+08:00 - root - INFO - Loading pretrained params from C:\Users\tom\Python_Project\vox-box\downloads\cache\model_scope\iic\speech_paraformer-large-vad-punc_asr_nat-zh-cn-16k-common-vocab8404-pytorch\model.pt
2025-06-12T16:44:42+08:00 - root - INFO - ckpt: C:\Users\tom\Python_Project\vox-box\downloads\cache\model_scope\iic\speech_paraformer-large-vad-punc_asr_nat-zh-cn-16k-common-vocab8404-pytorch\model.pt
2025-06-12T16:44:42+08:00 - root - INFO - scope_map: ['module.', 'None']
2025-06-12T16:44:42+08:00 - root - INFO - excludes: None
2025-06-12T16:44:42+08:00 - root - INFO - Loading ckpt: C:\Users\tom\Python_Project\vox-box\downloads\cache\model_scope\iic\speech_paraformer-large-vad-punc_asr_nat-zh-cn-16k-common-vocab8404-pytorch\model.pt, status:
2025-06-12T16:44:42+08:00 - asyncio - DEBUG - Using proactor: IocpProactor
2025-06-12T16:44:42+08:00 - vox_box.server.server - INFO - Starting Vox Box server.
2025-06-12T16:44:42+08:00 - vox_box.server.server - INFO - Serving on 0.0.0.0:8080.
2025-06-12T16:44:42+08:00 - httpx - DEBUG - load_ssl_context verify=True cert=None trust_env=True http2=False
2025-06-12T16:44:42+08:00 - httpx - DEBUG - load_verify_locations cafile='C:\Users\tom\AppData\Local\Programs\Python\Python311\Lib\site-packages\certifi\cacert.pem'
2025-06-12T16:44:57+08:00 - python_multipart.multipart - DEBUG - Calling on_part_begin with no data
2025-06-12T16:44:57+08:00 - python_multipart.multipart - DEBUG - Calling on_header_field with data[36:55]
2025-06-12T16:44:57+08:00 - python_multipart.multipart - DEBUG - Calling on_header_value with data[57:80]
2025-06-12T16:44:57+08:00 - python_multipart.multipart - DEBUG - Calling on_header_end with no data
2025-06-12T16:44:57+08:00 - python_multipart.multipart - DEBUG - Calling on_headers_finished with no data
2025-06-12T16:44:57+08:00 - python_multipart.multipart - DEBUG - Calling on_part_data with data[84:97]
2025-06-12T16:44:57+08:00 - python_multipart.multipart - DEBUG - Calling on_part_end with no data
2025-06-12T16:44:57+08:00 - python_multipart.multipart - DEBUG - Calling on_part_begin with no data
2025-06-12T16:44:57+08:00 - python_multipart.multipart - DEBUG - Calling on_header_field with data[135:154]
2025-06-12T16:44:57+08:00 - python_multipart.multipart - DEBUG - Calling on_header_value with data[156:231]
2025-06-12T16:44:57+08:00 - python_multipart.multipart - DEBUG - Calling on_header_end with no data
2025-06-12T16:44:57+08:00 - python_multipart.multipart - DEBUG - Calling on_headers_finished with no data
2025-06-12T16:44:57+08:00 - python_multipart.multipart - DEBUG - Calling on_part_data with data[235:13032]
2025-06-12T16:44:57+08:00 - python_multipart.multipart - DEBUG - Calling on_part_data with data[0:17008]
2025-06-12T16:44:57+08:00 - python_multipart.multipart - DEBUG - Calling on_part_end with no data
2025-06-12T16:44:57+08:00 - python_multipart.multipart - DEBUG - Calling on_end with no data
2025-06-12T16:44:57+08:00 - python_multipart.multipart - WARNING - Skipping data after last boundary

[Hunter6324]

Okay, here's a summary of the issues encountered and proposed solutions in English:

Current Situation Summary:

1. Goal: Enable OpenWebUI (running in Docker) to use vox-box (Paraformer model, running on the host machine) for speech-to-text (STT).
2. Network Connectivity: The OpenWebUI Docker container can successfully connect to the vox-box service on the host (http://host.docker.internal:8080).
3. OpenWebUI Request Flow:
   • The browser records audio.
   • The browser sends the audio as multipart/form-data to OpenWebUI's own backend API endpoint (http://localhost:10097/api/v1/audio/transcriptions).
   • The OpenWebUI backend is then supposed to forward this request to the configured vox-box service.
4. MIME Type Configuration:
   • The browser sends the audio file within the multipart/form-data with an internal Content-Type of audio/webm;codecs=opus.
   • Setting OpenWebUI's "Supported MIME Types" for STT to audio/webm;codecs=opus allows the OpenWebUI backend to accept the request from the frontend (it no longer immediately rejects due to MIME type mismatch).
5. Core Current Issue:
   • Requests from OpenWebUI now reach the vox-box service (evidenced by python-multipart DEBUG logs in vox-box).
   • However, vox-box processes this request and returns an HTTP 400 error.
   • OpenWebUI displays this as [ERROR: 400: [ERROR: 'text']].
   • Crucially, vox-box's application-level logs do not show a specific error почему it's returning the 400 error. Only low-level python-multipart parsing logs are visible.
6. Successful Direct vox-box Test:
   • Using a modified test_transcription.py script to directly send a .webm (Opus encoded) audio file and the paraformer-zh model parameter to vox-box results in successful transcription.
   • This proves vox-box itself can handle this audio format and model.

Inferred Root Cause:

Since the direct script call to vox-box works, but the request forwarded through OpenWebUI causes vox-box to return a 400 error (without clear application-level error logs from vox-box), the most probable cause is a mismatch between the structure of the multipart/form-data request forwarded by OpenWebUI and what vox-box expects.

The vox-box /v1/audio/transcriptions endpoint expects multipart/form-data containing specific fields. Based on the successful test_transcription.py script, it requires at least:

• A field named file containing the audio data.
• A field named model containing the model name (e.g., paraformer-zh).

If OpenWebUI, when forwarding the request, fails to correctly include these (or other necessary) fields, or if the field names don't match, vox-box's FastAPI application might fail when trying to bind these required parameters from the request. This kind of parameter binding failure might result in a 400 error without producing a detailed error stack trace at vox-box's default logging level.

Solution Suggestions & Next Steps:

1. Verify Form Fields Sent by OpenWebUI Frontend:
   This is the most critical next step. We need to confirm what form fields OpenWebUI's frontend is sending to its own backend (/api/v1/audio/transcriptions).

   • Action: In your browser's Developer Tools (Network tab), inspect the POST request made to OpenWebUI's backend (http://localhost:10097/api/v1/audio/transcriptions). Look at the "Payload" (or "Form Data") section.
   • Check:
     • Is there a part named file (or similar) containing the audio data? (We previously confirmed this exists).
     • Is there a part named model? Is its value paraformer-zh?
     • Are there any other potentially relevant form fields?

2. Act Based on the Findings from Step 1:

   • If the model field is missing from the Payload, or its value is incorrect:
     • The problem lies with OpenWebUI's frontend or its configuration. It's not correctly reading the "Speech-to-Text Model" setting (paraformer-zh) and including it in the request to its backend.
     • You'll need to double-check the relevant settings in OpenWebUI to ensure the model name is saved correctly. If the setting is correct but the field is still not sent, this could be a bug in OpenWebUI.
   • If both file and model fields are present in the Payload with correct values:
     • This means OpenWebUI's frontend is correctly sending the data to its backend. The issue then shifts to whether OpenWebUI's backend is accurately and completely forwarding these form fields when it makes the request to vox-box.
     • In this scenario, since vox-box isn't providing a clear application-level error:
       • Try to increase vox-box's logging verbosity: See if vox-box has options for more detailed logging or a debug mode that would show the exact request parameters it receives and its internal processing steps. This might involve changing vox-box startup parameters or its code.
       • Examine OpenWebUI Source Code (if feasible): Look at the OpenWebUI backend code responsible for handling /api/v1/audio/transcriptions and forwarding to a custom STT service. This would reveal how it constructs the request to vox-box.

For now, please proceed with Step 1 above: inspect the form data sent by the OpenWebUI frontend.