aliyun-qwen-asr

Name: aliyun-qwen-asr
Author: cinience

59installs34forks

How to install aliyun-qwen-asr?

cinience/alicloud-skills/aliyun-qwen-asr

$npx -y skills add cinience/alicloud-skills --skill aliyun-qwen-asr

Installs into the current project.

Summary

Use when transcribing non-realtime speech with Alibaba Cloud Model Studio Qwen ASR models (qwen3-asr-flash, qwen-audio-asr, qwen3-asr-flash-filetrans). Use when converting recorded audio files to text, generating transcripts with timestamps, or documenting DashScope/OpenAI-

Files · 1

View on GitHub

SKILL.md

1	Category: provider
2
3	# Model Studio Qwen ASR (Non-Realtime)
4
5	## Validation
6
7	```bash
8	mkdir -p output/aliyun-qwen-asr
9	python -m py_compile skills/ai/audio/aliyun-qwen-asr/scripts/transcribe_audio.py && echo "py_compile_ok" > output/aliyun-qwen-asr/validate.txt
10	```
11
12	Pass criteria: command exits 0 and `output/aliyun-qwen-asr/validate.txt` is generated.
13
14	## Output And Evidence
15
16	- Store transcripts and API responses under `output/aliyun-qwen-asr/`.
17	- Keep one command log or sample response per run.
18
19	Use Qwen ASR for recorded audio transcription (non-realtime), including short audio sync calls and long audio async jobs.
20
21	## Critical model names
22
23	Use one of these exact model strings:
24	- `qwen3-asr-flash`
25	- `qwen3-asr-flash-2026-02-10`
26	- `qwen-audio-asr`
27	- `qwen3-asr-flash-filetrans`
28	- `qwen3-asr-flash-filetrans-2025-11-17`
29
30	Selection guidance:
31	- Use `qwen3-asr-flash`, `qwen3-asr-flash-2026-02-10`, or `qwen-audio-asr` for short/normal recordings (sync).
32	- Use `qwen3-asr-flash-filetrans` or `qwen3-asr-flash-filetrans-2025-11-17` for long-file transcription (async task workflow).
33
34	## Prerequisites
35
36	- Install SDK dependencies (script uses Python stdlib only):
37
38	```bash
39	python3 -m venv .venv
40	. .venv/bin/activate
41	```
42
43	- Set `DASHSCOPE_API_KEY` in environment, or add `dashscope_api_key` to `~/.alibabacloud/credentials`.
44
45	## Normalized interface (asr.transcribe)
46
47	### Request
48	- `audio` (string, required): public URL or local file path.
49	- `model` (string, optional): default `qwen3-asr-flash`.
50	- `language_hints` (array<string>, optional): e.g. `zh`, `en`.
51	- `sample_rate` (number, optional)
52	- `vocabulary_id` (string, optional)
53	- `disfluency_removal_enabled` (bool, optional)
54	- `timestamp_granularities` (array<string>, optional): e.g. `sentence`.
55	- `async` (bool, optional): default false for sync models, true for `qwen3-asr-flash-filetrans`.
56
57	### Response
58	- `text` (string): normalized transcript text.
59	- `task_id` (string, optional): present for async submission.
60	- `status` (string): `SUCCEEDED` or submission status.
61	- `raw` (object): original API response.
62
63	## Quick start (official HTTP API)
64
65	Sync transcription (OpenAI-compatible protocol):
66
67	```bash
68	curl -sS --location 'https://dashscope.aliyuncs.com/compatible-mode/v1/chat/completions' \
69	--header "Authorization: Bearer $DASHSCOPE_API_KEY" \
70	--header 'Content-Type: application/json' \
71	--data '{
72	"model": "qwen3-asr-flash",
73	"messages": [
74	{
75	"role": "user",
76	"content": [
77	{
78	"type": "input_audio",
79	"input_audio": {
80	"data": "https://dashscope.oss-cn-beijing.aliyuncs.com/audios/welcome.mp3"
81	}
82	}
83	]
84	}
85	],
86	"stream": false,
87	"asr_options": {
88	"enable_itn": false
89	}
90	}'
91	```
92
93	Async long-file transcription (DashScope protocol):
94
95	```bash
96	curl -sS --location 'https://dashscope.aliyuncs.com/api/v1/services/audio/asr/transcription' \
97	--header "Authorization: Bearer $DASHSCOPE_API_KEY" \
98	--header 'X-DashScope-Async: enable' \
99	--header 'Content-Type: application/json' \
100	--data '{
101	"model": "qwen3-asr-flash-filetrans",
102	"input": {
103	"file_url": "https://dashscope.oss-cn-beijing.aliyuncs.com/audios/welcome.mp3"
104	}
105	}'
106	```
107
108	Poll task result:
109
110	```bash
111	curl -sS --location "https://dashscope.aliyuncs.com/api/v1/tasks/<task_id>" \
112	--header "Authorization: Bearer $DASHSCOPE_API_KEY"
113	```
114
115	## Local helper script
116
117	Use the bundled script for URL/local-file input and optional async polling:
118
119	```bash
120	python skills/ai/audio/aliyun-qwen-asr/scripts/transcribe_audio.py \
121	--audio "https://dashscope.oss-cn-beijing.aliyuncs.com/audios/welcome.mp3" \
122	--model qwen3-asr-flash \
123	--language-hints zh,en \
124	--print-response
125	```
126
127	Long-file mode:
128
129	```bash
130	python skills/ai/audio/aliyun-qwen-asr/scripts/transcribe_audio.py \
131	--audio "https://dashscope.oss-cn-beijing.aliyuncs.com/audios/welcome.mp3" \
132	--model qwen3-asr-flash-filetrans \
133	--async \
134	--wait
135	```
136
137	## Operational guidance
138
139	- For local files, use `input_audio.data` (data URI) when direct URL is unavailable.
140	- Keep `language_hints` minimal to reduce recognition ambiguity.
141	- For async tasks, use 5-20s polling interval with max retry guard.
142	- Save normalized outputs under `output/aliyun-qwen-asr/transcripts/`.
143
144	## Output location
145
146	- Default output: `output/aliyun-qwen-asr/transcripts/`
147	- Override base dir with `OUTPUT_DIR`.
148
149	## Workflow
150
151	1) Confirm user intent, region, identifiers, and whether the operation is read-only or mutating.
152	2) Run one minimal read-only query first to verify connectivity and permissions.
153	3) Execute the tar

People who liked this also viewed

6 picks

Type

aliyun-qwen-asr

59installs34forks

View on GitHub

How to install aliyun-qwen-asr?

cinience/alicloud-skills/aliyun-qwen-asr

$npx -y skills add cinience/alicloud-skills --skill aliyun-qwen-asr

Installs into the current project.

Summary

Files · 1

View on GitHub

SKILL.md

1	Category: provider
2
3	# Model Studio Qwen ASR (Non-Realtime)
4
5	## Validation
6
7	```bash
8	mkdir -p output/aliyun-qwen-asr
9	python -m py_compile skills/ai/audio/aliyun-qwen-asr/scripts/transcribe_audio.py && echo "py_compile_ok" > output/aliyun-qwen-asr/validate.txt
10	```
11
12	Pass criteria: command exits 0 and `output/aliyun-qwen-asr/validate.txt` is generated.
13
14	## Output And Evidence
15
16	- Store transcripts and API responses under `output/aliyun-qwen-asr/`.
17	- Keep one command log or sample response per run.
18
19	Use Qwen ASR for recorded audio transcription (non-realtime), including short audio sync calls and long audio async jobs.
20
21	## Critical model names
22
23	Use one of these exact model strings:
24	- `qwen3-asr-flash`
25	- `qwen3-asr-flash-2026-02-10`
26	- `qwen-audio-asr`
27	- `qwen3-asr-flash-filetrans`
28	- `qwen3-asr-flash-filetrans-2025-11-17`
29
30	Selection guidance:
31	- Use `qwen3-asr-flash`, `qwen3-asr-flash-2026-02-10`, or `qwen-audio-asr` for short/normal recordings (sync).
32	- Use `qwen3-asr-flash-filetrans` or `qwen3-asr-flash-filetrans-2025-11-17` for long-file transcription (async task workflow).
33
34	## Prerequisites
35
36	- Install SDK dependencies (script uses Python stdlib only):
37
38	```bash
39	python3 -m venv .venv
40	. .venv/bin/activate
41	```
42
43	- Set `DASHSCOPE_API_KEY` in environment, or add `dashscope_api_key` to `~/.alibabacloud/credentials`.
44
45	## Normalized interface (asr.transcribe)
46
47	### Request
48	- `audio` (string, required): public URL or local file path.
49	- `model` (string, optional): default `qwen3-asr-flash`.
50	- `language_hints` (array<string>, optional): e.g. `zh`, `en`.
51	- `sample_rate` (number, optional)
52	- `vocabulary_id` (string, optional)
53	- `disfluency_removal_enabled` (bool, optional)
54	- `timestamp_granularities` (array<string>, optional): e.g. `sentence`.
55	- `async` (bool, optional): default false for sync models, true for `qwen3-asr-flash-filetrans`.
56
57	### Response
58	- `text` (string): normalized transcript text.
59	- `task_id` (string, optional): present for async submission.
60	- `status` (string): `SUCCEEDED` or submission status.
61	- `raw` (object): original API response.
62
63	## Quick start (official HTTP API)
64
65	Sync transcription (OpenAI-compatible protocol):
66
67	```bash
68	curl -sS --location 'https://dashscope.aliyuncs.com/compatible-mode/v1/chat/completions' \
69	--header "Authorization: Bearer $DASHSCOPE_API_KEY" \
70	--header 'Content-Type: application/json' \
71	--data '{
72	"model": "qwen3-asr-flash",
73	"messages": [
74	{
75	"role": "user",
76	"content": [
77	{
78	"type": "input_audio",
79	"input_audio": {
80	"data": "https://dashscope.oss-cn-beijing.aliyuncs.com/audios/welcome.mp3"
81	}
82	}
83	]
84	}
85	],
86	"stream": false,
87	"asr_options": {
88	"enable_itn": false
89	}
90	}'
91	```
92
93	Async long-file transcription (DashScope protocol):
94
95	```bash
96	curl -sS --location 'https://dashscope.aliyuncs.com/api/v1/services/audio/asr/transcription' \
97	--header "Authorization: Bearer $DASHSCOPE_API_KEY" \
98	--header 'X-DashScope-Async: enable' \
99	--header 'Content-Type: application/json' \
100	--data '{
101	"model": "qwen3-asr-flash-filetrans",
102	"input": {
103	"file_url": "https://dashscope.oss-cn-beijing.aliyuncs.com/audios/welcome.mp3"
104	}
105	}'
106	```
107
108	Poll task result:
109
110	```bash
111	curl -sS --location "https://dashscope.aliyuncs.com/api/v1/tasks/<task_id>" \
112	--header "Authorization: Bearer $DASHSCOPE_API_KEY"
113	```
114
115	## Local helper script
116
117	Use the bundled script for URL/local-file input and optional async polling:
118
119	```bash
120	python skills/ai/audio/aliyun-qwen-asr/scripts/transcribe_audio.py \
121	--audio "https://dashscope.oss-cn-beijing.aliyuncs.com/audios/welcome.mp3" \
122	--model qwen3-asr-flash \
123	--language-hints zh,en \
124	--print-response
125	```
126
127	Long-file mode:
128
129	```bash
130	python skills/ai/audio/aliyun-qwen-asr/scripts/transcribe_audio.py \
131	--audio "https://dashscope.oss-cn-beijing.aliyuncs.com/audios/welcome.mp3" \
132	--model qwen3-asr-flash-filetrans \
133	--async \
134	--wait
135	```
136
137	## Operational guidance
138
139	- For local files, use `input_audio.data` (data URI) when direct URL is unavailable.
140	- Keep `language_hints` minimal to reduce recognition ambiguity.
141	- For async tasks, use 5-20s polling interval with max retry guard.
142	- Save normalized outputs under `output/aliyun-qwen-asr/transcripts/`.
143
144	## Output location
145
146	- Default output: `output/aliyun-qwen-asr/transcripts/`
147	- Override base dir with `OUTPUT_DIR`.
148
149	## Workflow
150
151	1) Confirm user intent, region, identifiers, and whether the operation is read-only or mutating.
152	2) Run one minimal read-only query first to verify connectivity and permissions.
153	3) Execute the tar

People who liked this also viewed

6 picks

Type