dashscope

Name: dashscope
Author: calesthio

95installs4.9kforks

How to install dashscope?

calesthio/openmontage/dashscope

$npx -y skills add calesthio/OpenMontage --skill dashscope

Installs into the current project.

Summary

DashScope (Alibaba Cloud Bailian / 阿里云百炼) integration — image generation (qwen-image-2.0-pro), text-to-speech (qwen3-tts-flash), and ASR with word-level timestamps (qwen3-asr-flash-filetrans). Use when generating images via Qwen-Image, narrating via Qwen-TTS, or transcribing with

Files · 1

View on GitHub

SKILL.md

1	# DashScope
2
3	Requires `DASHSCOPE_API_KEY` in `.env`. Get one at https://dashscope.aliyun.com/.
4
5	## Current API
6
7	CRITICAL: DashScope's `/compatible-mode/v1/` only supports `/chat/completions` and `/embeddings`. Image generation, TTS, and ASR all use DashScope-native endpoints — not OpenAI-compatible paths.
8
9	All three tools use `Authorization: Bearer $DASHSCOPE_API_KEY`.
10
11	### Image Generation
12
13	```text
14	POST https://dashscope.aliyuncs.com/api/v1/services/aigc/multimodal-generation/generation
15	```
16
17	- Model: `qwen-image-2.0-pro` (default), `qwen-image-max`, `wan2.7-image`, `z-image-turbo`
18	- Body: `{model, input: {messages: [{role: "user", content: [{text: "prompt"}]}]}, parameters: {size: "W*H", n, prompt_extend, watermark}}`
19	- Size format uses asterisk: `"1024*1024"` not `"1024x1024"`
20	- Response: `output.choices[0].message.content[0].image` (URL, valid ~24h) — must download separately
21
22	### Text-to-Speech
23
24	```text
25	POST https://dashscope.aliyuncs.com/api/v1/services/aigc/multimodal-generation/generation
26	```
27
28	Same endpoint as image gen, different body.
29
30	- Model: `qwen3-tts-flash` (default), `qwen3-tts-instruct-flash`, `qwen-tts-2025-05-22`
31	- Body: `{model, input: {text, voice: "Cherry", language_type: "Auto"}}`
32	- Response: `output.audio.url` (WAV, valid ~24h) — must download separately
33
34	### ASR with Word-Level Timestamps
35
36	```text
37	POST https://dashscope.aliyuncs.com/api/v1/services/audio/asr/transcription
38	Header: X-DashScope-Async: enable
39	```
40
41	- Model: `qwen3-asr-flash-filetrans` (NOT `qwen3-asr-flash` — the sync version has no word timestamps)
42	- Body: `{model, input: {file_url: "https://public-url/audio.mp3"}, parameters: {enable_words: true, language_hints: ["zh","en"]}}`
43	- Returns `task_id` → poll `GET /api/v1/tasks/{task_id}` until `SUCCEEDED` → download `output.result.transcription_url` → JSON with `transcripts[].sentences[].words[]`
44	- Timestamps in `begin_time`/`end_time` are in milliseconds — the tool normalizes to seconds
45
46	## OpenMontage Usage
47
48	### Image via selector
49
50	```python
51	from tools.graphics.image_selector import ImageSelector
52
53	result = ImageSelector().execute({
54	"preferred_provider": "dashscope",
55	"prompt": "一只猫坐在沙发上",
56	"output_path": "projects/my-video/assets/images/cat.png",
57	})
58	```
59
60	### TTS via selector
61
62	```python
63	from tools.audio.tts_selector import TTSSelector
64
65	result = TTSSelector().execute({
66	"preferred_provider": "dashscope",
67	"text": "如果 AI 真的会改变未来，普通人到底该怎么参与？",
68	"voice": "Cherry",
69	"output_path": "projects/my-video/assets/audio/narration.wav",
70	})
71	```
72
73	### ASR directly (word timestamps for subtitles)
74
75	```python
76	from tools.analysis.dashscope_asr import DashscopeAsr
77
78	result = DashscopeAsr().execute({
79	"audio_url": "https://example.com/narration.wav",
80	"output_path": "projects/my-video/assets/audio/transcription.json",
81	})
82
83	# result.data["words"] is a flat list of {text, begin_time_seconds, end_time_seconds}
84	```
85
86	## Recommended Workflow
87
88	1. Image: Generate a sample first. Check `prompt_extend: true` (default) — DashScope rewrites your prompt for better results. Disable if you need literal prompt adherence.
89	2. TTS: Generate a 10-15 second sample before full narration. Approve voice and pacing before committing to full generation.
90	3. ASR: Audio must be at a publicly accessible URL. Upload to any public host (S3, etc.) first. Local paths are rejected with a clear error.
91	4. Subtitles: Build from `result.data["words"]` — each word has `begin_time_seconds` and `end_time_seconds`. Group words into caption phrases by language semantics, not fixed character count.
92
93	## Parameters
94
95	### Image (`dashscope_image`)
96	- `prompt` (required): text prompt
97	- `model`: default `qwen-image-2.0-pro`
98	- `size`: default `"10241024"` — asterisk separator, not "x"*
99	- `n`: 1-6 images
100	- `negative_prompt`: things to avoid (max 500 chars)
101	- `prompt_extend`: default `true` — auto-rewrite prompt for better results
102	- `watermark`: default `false`
103	- `seed`: for reproducibility
104
105	### TTS (`dashscope_tts`)
106	- `text` (required): text to synthesize (max 600 chars for qwen3-tts-flash)
107	- `model`: default `qwen3-tts-flash`
108	- `voice`: default `"Cherry"` — other voices: `"Ethan"`, `"Chelsie"`, etc.
109	- `language_type`: default `"Auto"` — `"Chinese"`, `"English"`, `"Japanese"`, `"Korean"`
110	- `instructions`: natural language delivery instructions (only for `qwen3-tts-instruct-flash`)
111
112	### ASR (`dashscope_asr`)
113	- `audio_url` (required): must be publicly accessible URL
114	- `model`: `qwen3-asr-flash-filetrans` (only model that supports word timestamps)
115	- `language_hints`: default `["zh", "en"]`
116	- `enable_words`: default `true` — required fo

People who liked this also viewed

6 picks

Type

dashscope

95installs4.9kforks

View on GitHub

How to install dashscope?

calesthio/openmontage/dashscope

$npx -y skills add calesthio/OpenMontage --skill dashscope

Installs into the current project.

Summary

Files · 1

View on GitHub

SKILL.md

1	# DashScope
2
3	Requires `DASHSCOPE_API_KEY` in `.env`. Get one at https://dashscope.aliyun.com/.
4
5	## Current API
6
7	CRITICAL: DashScope's `/compatible-mode/v1/` only supports `/chat/completions` and `/embeddings`. Image generation, TTS, and ASR all use DashScope-native endpoints — not OpenAI-compatible paths.
8
9	All three tools use `Authorization: Bearer $DASHSCOPE_API_KEY`.
10
11	### Image Generation
12
13	```text
14	POST https://dashscope.aliyuncs.com/api/v1/services/aigc/multimodal-generation/generation
15	```
16
17	- Model: `qwen-image-2.0-pro` (default), `qwen-image-max`, `wan2.7-image`, `z-image-turbo`
18	- Body: `{model, input: {messages: [{role: "user", content: [{text: "prompt"}]}]}, parameters: {size: "W*H", n, prompt_extend, watermark}}`
19	- Size format uses asterisk: `"1024*1024"` not `"1024x1024"`
20	- Response: `output.choices[0].message.content[0].image` (URL, valid ~24h) — must download separately
21
22	### Text-to-Speech
23
24	```text
25	POST https://dashscope.aliyuncs.com/api/v1/services/aigc/multimodal-generation/generation
26	```
27
28	Same endpoint as image gen, different body.
29
30	- Model: `qwen3-tts-flash` (default), `qwen3-tts-instruct-flash`, `qwen-tts-2025-05-22`
31	- Body: `{model, input: {text, voice: "Cherry", language_type: "Auto"}}`
32	- Response: `output.audio.url` (WAV, valid ~24h) — must download separately
33
34	### ASR with Word-Level Timestamps
35
36	```text
37	POST https://dashscope.aliyuncs.com/api/v1/services/audio/asr/transcription
38	Header: X-DashScope-Async: enable
39	```
40
41	- Model: `qwen3-asr-flash-filetrans` (NOT `qwen3-asr-flash` — the sync version has no word timestamps)
42	- Body: `{model, input: {file_url: "https://public-url/audio.mp3"}, parameters: {enable_words: true, language_hints: ["zh","en"]}}`
43	- Returns `task_id` → poll `GET /api/v1/tasks/{task_id}` until `SUCCEEDED` → download `output.result.transcription_url` → JSON with `transcripts[].sentences[].words[]`
44	- Timestamps in `begin_time`/`end_time` are in milliseconds — the tool normalizes to seconds
45
46	## OpenMontage Usage
47
48	### Image via selector
49
50	```python
51	from tools.graphics.image_selector import ImageSelector
52
53	result = ImageSelector().execute({
54	"preferred_provider": "dashscope",
55	"prompt": "一只猫坐在沙发上",
56	"output_path": "projects/my-video/assets/images/cat.png",
57	})
58	```
59
60	### TTS via selector
61
62	```python
63	from tools.audio.tts_selector import TTSSelector
64
65	result = TTSSelector().execute({
66	"preferred_provider": "dashscope",
67	"text": "如果 AI 真的会改变未来，普通人到底该怎么参与？",
68	"voice": "Cherry",
69	"output_path": "projects/my-video/assets/audio/narration.wav",
70	})
71	```
72
73	### ASR directly (word timestamps for subtitles)
74
75	```python
76	from tools.analysis.dashscope_asr import DashscopeAsr
77
78	result = DashscopeAsr().execute({
79	"audio_url": "https://example.com/narration.wav",
80	"output_path": "projects/my-video/assets/audio/transcription.json",
81	})
82
83	# result.data["words"] is a flat list of {text, begin_time_seconds, end_time_seconds}
84	```
85
86	## Recommended Workflow
87
88	1. Image: Generate a sample first. Check `prompt_extend: true` (default) — DashScope rewrites your prompt for better results. Disable if you need literal prompt adherence.
89	2. TTS: Generate a 10-15 second sample before full narration. Approve voice and pacing before committing to full generation.
90	3. ASR: Audio must be at a publicly accessible URL. Upload to any public host (S3, etc.) first. Local paths are rejected with a clear error.
91	4. Subtitles: Build from `result.data["words"]` — each word has `begin_time_seconds` and `end_time_seconds`. Group words into caption phrases by language semantics, not fixed character count.
92
93	## Parameters
94
95	### Image (`dashscope_image`)
96	- `prompt` (required): text prompt
97	- `model`: default `qwen-image-2.0-pro`
98	- `size`: default `"10241024"` — asterisk separator, not "x"*
99	- `n`: 1-6 images
100	- `negative_prompt`: things to avoid (max 500 chars)
101	- `prompt_extend`: default `true` — auto-rewrite prompt for better results
102	- `watermark`: default `false`
103	- `seed`: for reproducibility
104
105	### TTS (`dashscope_tts`)
106	- `text` (required): text to synthesize (max 600 chars for qwen3-tts-flash)
107	- `model`: default `qwen3-tts-flash`
108	- `voice`: default `"Cherry"` — other voices: `"Ethan"`, `"Chelsie"`, etc.
109	- `language_type`: default `"Auto"` — `"Chinese"`, `"English"`, `"Japanese"`, `"Korean"`
110	- `instructions`: natural language delivery instructions (only for `qwen3-tts-instruct-flash`)
111
112	### ASR (`dashscope_asr`)
113	- `audio_url` (required): must be publicly accessible URL
114	- `model`: `qwen3-asr-flash-filetrans` (only model that supports word timestamps)
115	- `language_hints`: default `["zh", "en"]`
116	- `enable_words`: default `true` — required fo

People who liked this also viewed

6 picks

Type