vision
kototTurn images into answers you can use.
image understanding
What it does
Delegates image understanding to a configurable OpenAI-compatible vision model and returns a text answer.
When to use it
Use it whenever a task requires understanding image content that the agent cannot view directly, including OCR, screenshots, diagrams, charts, and scanned documents.
How to use it
Give it an image path or URL and a clear question; it returns the vision model's text answer.
What you provide
- an image path or URL
- a specific question
- Your files
Access · 3
This skill
OpenAI-compatible vision endpoint
Sends images to vision endpoint
.claude/settings.json files
Reads vision configuration files
OpenClaw dotenv files
Reads OpenClaw configuration
What you need · 3
Python 3 must be available to run the script.
A configured OpenAI-compatible vision model endpoint must be available; the base URL and model are required.
An API key can be supplied through the environment or settings configuration; it is optional for local servers.
About this skill
- Visibility
- Public
- Repository
- kotot/vision
- Created
- Oct 8, 2026
- Updated
- Oct 8, 2026
- Files
- 2