ai-multimodal_mrgoonie

Name: ai-multimodal_mrgoonie
Author: jackspace

15stars7forks

How to install ai-multimodal_mrgoonie?

jackspace/claudeskillz/ai-multimodal_mrgoonie

$npx -y skills add jackspace/ClaudeSkillz --skill ai-multimodal_mrgoonie

Installs into the current project.

Summary

Process and generate multimedia content using Google Gemini API. Capabilities include analyze audio files (transcription with timestamps, summarization, speech understanding, music/sound analysis up to 9.5 hours), understand images (captioning, object detection, OCR, visual Q&A,

Files · 1

View on GitHub

SKILL.md

1	# AI Multimodal Processing Skill
2
3	Process audio, images, videos, documents, and generate images using Google Gemini's multimodal API. Unified interface for all multimedia content understanding and generation.
4
5	## Core Capabilities
6
7	### Audio Processing
8	- Transcription with timestamps (up to 9.5 hours)
9	- Audio summarization and analysis
10	- Speech understanding and speaker identification
11	- Music and environmental sound analysis
12	- Text-to-speech generation with controllable voice
13
14	### Image Understanding
15	- Image captioning and description
16	- Object detection with bounding boxes (2.0+)
17	- Pixel-level segmentation (2.5+)
18	- Visual question answering
19	- Multi-image comparison (up to 3,600 images)
20	- OCR and text extraction
21
22	### Video Analysis
23	- Scene detection and summarization
24	- Video Q&A with temporal understanding
25	- Transcription with visual descriptions
26	- YouTube URL support
27	- Long video processing (up to 6 hours)
28	- Frame-level analysis
29
30	### Document Extraction
31	- Native PDF vision processing (up to 1,000 pages)
32	- Table and form extraction
33	- Chart and diagram analysis
34	- Multi-page document understanding
35	- Structured data output (JSON schema)
36	- Format conversion (PDF to HTML/JSON)
37
38	### Image Generation
39	- Text-to-image generation
40	- Image editing and modification
41	- Multi-image composition (up to 3 images)
42	- Iterative refinement
43	- Multiple aspect ratios (1:1, 16:9, 9:16, 4:3, 3:4)
44	- Controllable style and quality
45
46	## Capability Matrix
47
48	\| Task \| Audio \| Image \| Video \| Document \| Generation \|
49	\|------\|:-----:\|:-----:\|:-----:\|:--------:\|:----------:\|
50	\| Transcription \| ✓ \| - \| ✓ \| - \| - \|
51	\| Summarization \| ✓ \| ✓ \| ✓ \| ✓ \| - \|
52	\| Q&A \| ✓ \| ✓ \| ✓ \| ✓ \| - \|
53	\| Object Detection \| - \| ✓ \| ✓ \| - \| - \|
54	\| Text Extraction \| - \| ✓ \| - \| ✓ \| - \|
55	\| Structured Output \| ✓ \| ✓ \| ✓ \| ✓ \| - \|
56	\| Creation \| TTS \| - \| - \| - \| ✓ \|
57	\| Timestamps \| ✓ \| - \| ✓ \| - \| - \|
58	\| Segmentation \| - \| ✓ \| - \| - \| - \|
59
60	## Model Selection Guide
61
62	### Gemini 2.5 Series (Recommended)
63	- gemini-2.5-pro: Highest quality, all features, 1M-2M context
64	- gemini-2.5-flash: Best balance, all features, 1M-2M context
65	- gemini-2.5-flash-lite: Lightweight, segmentation support
66	- gemini-2.5-flash-image: Image generation only
67
68	### Gemini 2.0 Series
69	- gemini-2.0-flash: Fast processing, object detection
70	- gemini-2.0-flash-lite: Lightweight option
71
72	### Feature Requirements
73	- Segmentation: Requires 2.5+ models
74	- Object Detection: Requires 2.0+ models
75	- Multi-video: Requires 2.5+ models
76	- Image Generation: Requires flash-image model
77
78	### Context Windows
79	- 2M tokens: ~6 hours video (low-res) or ~2 hours (default)
80	- 1M tokens: ~3 hours video (low-res) or ~1 hour (default)
81	- Audio: 32 tokens/second (1 min = 1,920 tokens)
82	- PDF: 258 tokens/page (fixed)
83	- Image: 258-1,548 tokens based on size
84
85	## Quick Start
86
87	### Prerequisites
88
89	API Key Setup: Supports both Google AI Studio and Vertex AI.
90
91	The skill checks for `GEMINI_API_KEY` in this order:
92	1. Process environment: `export GEMINI_API_KEY="your-key"`
93	2. Project root: `.env`
94	3. `.claude/.env`
95	4. `.claude/skills/.env`
96	5. `.claude/skills/ai-multimodal/.env`
97
98	Get API key: https://aistudio.google.com/apikey
99
100	For Vertex AI:
101	```bash
102	export GEMINI_USE_VERTEX=true
103	export VERTEX_PROJECT_ID=your-gcp-project-id
104	export VERTEX_LOCATION=us-central1 # Optional
105	```
106
107	Install SDK:
108	```bash
109	pip install google-genai python-dotenv pillow
110	```
111
112	### Common Patterns
113
114	Transcribe Audio:
115	```bash
116	python scripts/gemini_batch_process.py \
117	--files audio.mp3 \
118	--task transcribe \
119	--model gemini-2.5-flash
120	```
121
122	Analyze Image:
123	```bash
124	python scripts/gemini_batch_process.py \
125	--files image.jpg \
126	--task analyze \
127	--prompt "Describe this image" \
128	--model gemini-2.5-flash
129	```
130
131	Process Video:
132	```bash
133	python scripts/gemini_batch_process.py \
134	--files video.mp4 \
135	--task analyze \
136	--prompt "Summarize key points with timestamps" \
137	--model gemini-2.5-flash
138	```
139
140	Extract from PDF:
141	```bash
142	python scripts/gemini_batch_process.py \
143	--files document.pdf \
144	--task extract \

People who liked this also viewed

6 picks

Type

ai-multimodal_mrgoonie

15stars7forks

View on GitHub

How to install ai-multimodal_mrgoonie?

jackspace/claudeskillz/ai-multimodal_mrgoonie

$npx -y skills add jackspace/ClaudeSkillz --skill ai-multimodal_mrgoonie

Installs into the current project.

Summary

Files · 1

View on GitHub

SKILL.md

1	# AI Multimodal Processing Skill
2
3	Process audio, images, videos, documents, and generate images using Google Gemini's multimodal API. Unified interface for all multimedia content understanding and generation.
4
5	## Core Capabilities
6
7	### Audio Processing
8	- Transcription with timestamps (up to 9.5 hours)
9	- Audio summarization and analysis
10	- Speech understanding and speaker identification
11	- Music and environmental sound analysis
12	- Text-to-speech generation with controllable voice
13
14	### Image Understanding
15	- Image captioning and description
16	- Object detection with bounding boxes (2.0+)
17	- Pixel-level segmentation (2.5+)
18	- Visual question answering
19	- Multi-image comparison (up to 3,600 images)
20	- OCR and text extraction
21
22	### Video Analysis
23	- Scene detection and summarization
24	- Video Q&A with temporal understanding
25	- Transcription with visual descriptions
26	- YouTube URL support
27	- Long video processing (up to 6 hours)
28	- Frame-level analysis
29
30	### Document Extraction
31	- Native PDF vision processing (up to 1,000 pages)
32	- Table and form extraction
33	- Chart and diagram analysis
34	- Multi-page document understanding
35	- Structured data output (JSON schema)
36	- Format conversion (PDF to HTML/JSON)
37
38	### Image Generation
39	- Text-to-image generation
40	- Image editing and modification
41	- Multi-image composition (up to 3 images)
42	- Iterative refinement
43	- Multiple aspect ratios (1:1, 16:9, 9:16, 4:3, 3:4)
44	- Controllable style and quality
45
46	## Capability Matrix
47
48	\| Task \| Audio \| Image \| Video \| Document \| Generation \|
49	\|------\|:-----:\|:-----:\|:-----:\|:--------:\|:----------:\|
50	\| Transcription \| ✓ \| - \| ✓ \| - \| - \|
51	\| Summarization \| ✓ \| ✓ \| ✓ \| ✓ \| - \|
52	\| Q&A \| ✓ \| ✓ \| ✓ \| ✓ \| - \|
53	\| Object Detection \| - \| ✓ \| ✓ \| - \| - \|
54	\| Text Extraction \| - \| ✓ \| - \| ✓ \| - \|
55	\| Structured Output \| ✓ \| ✓ \| ✓ \| ✓ \| - \|
56	\| Creation \| TTS \| - \| - \| - \| ✓ \|
57	\| Timestamps \| ✓ \| - \| ✓ \| - \| - \|
58	\| Segmentation \| - \| ✓ \| - \| - \| - \|
59
60	## Model Selection Guide
61
62	### Gemini 2.5 Series (Recommended)
63	- gemini-2.5-pro: Highest quality, all features, 1M-2M context
64	- gemini-2.5-flash: Best balance, all features, 1M-2M context
65	- gemini-2.5-flash-lite: Lightweight, segmentation support
66	- gemini-2.5-flash-image: Image generation only
67
68	### Gemini 2.0 Series
69	- gemini-2.0-flash: Fast processing, object detection
70	- gemini-2.0-flash-lite: Lightweight option
71
72	### Feature Requirements
73	- Segmentation: Requires 2.5+ models
74	- Object Detection: Requires 2.0+ models
75	- Multi-video: Requires 2.5+ models
76	- Image Generation: Requires flash-image model
77
78	### Context Windows
79	- 2M tokens: ~6 hours video (low-res) or ~2 hours (default)
80	- 1M tokens: ~3 hours video (low-res) or ~1 hour (default)
81	- Audio: 32 tokens/second (1 min = 1,920 tokens)
82	- PDF: 258 tokens/page (fixed)
83	- Image: 258-1,548 tokens based on size
84
85	## Quick Start
86
87	### Prerequisites
88
89	API Key Setup: Supports both Google AI Studio and Vertex AI.
90
91	The skill checks for `GEMINI_API_KEY` in this order:
92	1. Process environment: `export GEMINI_API_KEY="your-key"`
93	2. Project root: `.env`
94	3. `.claude/.env`
95	4. `.claude/skills/.env`
96	5. `.claude/skills/ai-multimodal/.env`
97
98	Get API key: https://aistudio.google.com/apikey
99
100	For Vertex AI:
101	```bash
102	export GEMINI_USE_VERTEX=true
103	export VERTEX_PROJECT_ID=your-gcp-project-id
104	export VERTEX_LOCATION=us-central1 # Optional
105	```
106
107	Install SDK:
108	```bash
109	pip install google-genai python-dotenv pillow
110	```
111
112	### Common Patterns
113
114	Transcribe Audio:
115	```bash
116	python scripts/gemini_batch_process.py \
117	--files audio.mp3 \
118	--task transcribe \
119	--model gemini-2.5-flash
120	```
121
122	Analyze Image:
123	```bash
124	python scripts/gemini_batch_process.py \
125	--files image.jpg \
126	--task analyze \
127	--prompt "Describe this image" \
128	--model gemini-2.5-flash
129	```
130
131	Process Video:
132	```bash
133	python scripts/gemini_batch_process.py \
134	--files video.mp4 \
135	--task analyze \
136	--prompt "Summarize key points with timestamps" \
137	--model gemini-2.5-flash
138	```
139
140	Extract from PDF:
141	```bash
142	python scripts/gemini_batch_process.py \
143	--files document.pdf \
144	--task extract \

People who liked this also viewed

6 picks

Type