skills.rest
goodoleusa
goodoleusaCOMMUNITY
Explore all skills by this author →
GitHub

vision-ingest

Extract structured data from screen recordings and UI screenshots into merged JSON.

Updated Mar 15, 2026
Search Tags:#deduplication#data-extraction#prisma-studio#frame-processing#vision-ingest#json-merge
One-click install
npx skills add https://github.com/goodoleusa/cybertemplate-pipeline --skill vision-ingest
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: vision-ingest
Source: https://github.com/goodoleusa/cybertemplate-pipeline/tree/main/.claude/skills/vision-ingest
Command: npx skills add https://github.com/goodoleusa/cybertemplate-pipeline --skill vision-ingest

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires anthropic, Pillow, and includes scripts (resource) components.

What problem does it solve?

The Vision Ingest pipeline automates the extraction of structured data from screen recordings or screenshot directories, converting frames into JSON for downstream analytics and dashboards.

Core Features & Use Cases

  • End-to-end pipeline: from raw frames to merged JSON with deduplication and frame provenance.
  • Local-first processing: all steps 1-3 are offline and cost-free, with optional Vision API processing for richer extractions.
  • Use Case: turn Prisma Studio screenshots into tabular data for dashboards, audits, and QA reports.

Quick Start

Place raw frames in the input directory and run the Vision Ingest pipeline to produce the final merged JSON.

Frequently Asked Questions about vision-ingest

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I extract structured data from UI screenshots into JSON?▼

Extracting structured data from UI screenshots into JSON requires automating frame extraction, deduplication, and Vision API processing. This pipeline converts screen recordings into a unified, deduplicated JSON dataset with frame provenance for downstream analytics.

What is the best way to convert Prisma Studio screen recordings into tabular data?▼

Converting Prisma Studio screen recordings into tabular data involves an end-to-end pipeline that extracts frames, compresses them, and merges the per-frame Vision API extractions. It outputs a final deduplicated JSON dataset suitable for dashboards and audits.

Do I need an Anthropic API key to process screen recording frames?▼

You do not strictly need an Anthropic API key for the initial local-first processing steps like frame extraction, deduplication, and compression, which are offline and cost-free. The Anthropic Vision API is optional for running richer per-frame data extractions.

Can I use Pillow for image compression before the Vision API extraction?▼

Yes, Pillow handles the local image compression and frame processing steps offline. This prepares the extracted UI frames locally before sending them to the Anthropic Vision API for structured data extraction and final JSON merging.

How does frame deduplication work when merging screen recording data?▼

Frame deduplication works by identifying and removing redundant UI frames during the local-first processing phase before Vision API extraction. This ensures the final merged JSON dataset contains only unique structured data points with accurate frame provenance.

Are there limitations to local-first processing for QA pipeline screenshots?▼

The limitation of local-first processing for QA pipeline screenshots is that offline steps only handle frame extraction and deduplication. Achieving richer structured data extraction requires the Anthropic Vision API, meaning full end-to-end automation depends on online API availability.

Related Skills

Explore More Data & Analytics Skills→
mazicimertmazicimert

screenshot-automation

Generate localized App Store screenshots with device frames via UI tests.

Community
Advanced
bcmcpherbcmcpher

gen-data-dict

Generate a JSON data dictionary describing variables in merged CSV, TSV, or Parquet datasets.

Community
Intermediate
kevinreberkevinreber
3

database-operation

Create Prisma database queries and server functions with typed responses.

Community
Advanced
wulaosijiwulaosiji
28

image-ocr

Extract text from images using PaddleOCR with Chinese and English support.

Community
Intermediate