AGENT SKILL
openai-vision
Analyze images and multi-frame sequences using OpenAI GPT vision models
- License 许可
- Apache-2.0
- 最近更新
- 2026-11-14
- Stars
- ★ 0
- Forks
- 1
01 前言 Front Matter
- name
- openai-vision
- description
- Analyze images and multi-frame sequences using OpenAI GPT vision models
03 SKILL.md 正文
(unchanged — see original skill content)
Steps
- Encode local images to base64 with correct media type, or pass URLs directly.
- Call
client.chat.completions.createwithmodel="gpt-4o"and acontentarray mixingtextandimage_urlparts. - Choose
detail:highfor OCR / small objects,lowfor scene classification,autootherwise. - For structured output, instruct the model to return only JSON and parse it, stripping code fences.
- Handle rate limits with exponential backoff; resize images over ~15MB before sending.
Expected Result
A parsed analysis (string or JSON) plus token-usage metadata for each processed image.
04 同仓库更多 Skills
video-frame-extraction
Extract frames from video files and save them as images using OpenCV
★ 0
openai-vision
Analyze images and multi-frame sequences using OpenAI GPT vision models
★ 0
image-ocr
Extract text content from images using Tesseract OCR via Python, including structured field extraction from receipts and forms
★ 0
video-frame-extraction
Extract frames from video files and save them as images using OpenCV
★ 0
openai-vision
Analyze images and multi-frame sequences using OpenAI GPT vision models
★ 0
image-ocr
Extract text content from images using Tesseract OCR via Python, including structured field extraction from receipts and forms
★ 0