Skills
VisoWork skills are self-contained capability bundles your AI agent loads on demand. Each skill wraps one or more VisoWork vision APIs — describe, OCR, detect, vector search — into an agent-native format: tool definitions, executable scripts, prompts, and runnable examples.
Skills execute inside VisoWork's sandbox, so your agent can call them without handling raw API credentials or writing integration code.
Organized by Domain
The skill catalog groups skills by problem domain. Pick the tag that matches your use case, or stay on General for domain-agnostic tools.
- General — Cross-cutting tools: describe, OCR, detect, vector search.
- Retail — Shelf audits, planogram checks, product recognition.
- Manufacturing — Defect detection, assembly verification, QA.
- Healthcare — Medical imaging aids, lab document parsing.
- Logistics — Package tracking, label reading, damage inspection.
- Education — Handwriting recognition, worksheet parsing, content tagging.
- Finance — Statement OCR, KYC document checks, chart extraction.
- Security — Surveillance analysis, access-control image review.
- Robotics — Scene understanding, obstacle recognition, visual grounding.
- Science — Chart and diagram interpretation, specimen annotation.
- Energy — Meter reading, infrastructure inspection, asset monitoring.
The in-app catalog filters use this same domain taxonomy — start broad, then narrow down.
How to Use
- Follow the Setup Guide to connect your agent — install the skill package directly, or wire it up over MCP.
- Browse the Catalog (or open /skills in the app) and pick the domain that matches your work.
- Open any skill to see its tools, inputs, and example prompts, then invoke it from your agent in natural language.
No SDK is required — once a skill is loaded, your agent selects the right tool from the user's request.