uu889's local workbench connecting OpenCV vision with decision models: OpenCV segments, detects, and measures; hard rules decide anything numeric; and a Laya or Jev model answers typed questions, with conflicts flagged for human review.

How does it involve Jev?

It treats Laya and TypeSafe's Jev as the same kind of decision model and can send the same request to a local model, TypeSafe's official Jev, aiask.me network acceleration, or any third-party /v1/systemone endpoint. Its training page is pitched as an "open-source JEV".

Download a Release: on Windows unzip and run install.bat then start.bat; on Linux run tar -xzf, then bash install.sh and bash start.sh, which opens the workbench at 127.0.0.1:8090. Python 3.10 or newer is required; Debian and Ubuntu need python3-venv installed first.

What can I process with it?

Five built-in detection schemes — fruit picking, weed removal, pest monitoring, steel-pipe QC, and textile QC — plus region-classification training and a general workbench where states can be text, JSON objects, or message lists with yes/no, single-choice, and score questions.

Can I train my own decision model?

Yes: choose from 11 built-in encoders or any Hugging Face model id, train in full, frozen-embedding, or LoRA mode with automatic downshifting on low VRAM, resume from checkpoints, evaluate on a holdout set, and set the result as the default model.

Do I need a GPU to train?

Not necessarily. The README says small models train on 4 GB of VRAM and that CPU training works without a graphics card; the decision-training screenshot in the docs was taken on a machine with no GPU using demo data.

Is it open source, and where do API keys go?

MIT-licensed: 57 stars, 8 forks, created 2026-10-05, last pushed 2026-10-08. Keys are added in the page and stored only on your machine; remote failures switch to another interface automatically, and batch requests split into per-item calls.

Jev directory

CommunityJev Tools & CLIs

Laya OpenCV

A local workbench that pairs OpenCV vision with Laya or Jev decision models: images are segmented and measured, hard rules settle the numbers, and the model gives the verdict — with one-click installers and an "open-source JEV" training path served at /v1/systemone.

See highlights Open github.com

Stars: 57
Language: Python
License: MIT
Reading time: 3 min
Published by: Community
Author: uu889
Last verified: Oct 2026

uu889/laya-opencv

User repository on GitHub

View on GitHub

OpenCV 与判定模型（Laya / Jev）结合的本地工作台：OpenCV 看图测量，硬规则管数字，模型下结论。用于水果采摘、杂草与害虫监测、钢管和纺织物质检。 | OpenCV + Laya/Jev decision models in a one-click local workbench: OpenCV measures, hard rules settle the numbers, the model gives the verdict. For agriculture and quality inspection.

Forks: 8
Last pushed: Updated Oct 2026

Repo stats from the GitHub API, cached Oct 2026.

Tagscommunity github python opencv quality-inspection

Works forJev for Classification Jev for Scoring Jev for Verification

Used inJev for Software & DevTools

Languages: Python
Platforms: API
Capabilities: Open sourceSelf-hostedRuns locally

Highlights

README: OpenCV measures the image, hard rules settle the numbers, and the decision model gives the verdict, with cross-checking and human review on conflicts.
3.0 adds “open-source JEV” training: build a Laya-format model from any Hugging Face encoder or fine-tune official weights, served at /v1/systemone.
Small models train on 4 GB of VRAM and CPU training works without a GPU, with full, frozen-embedding, and LoRA modes plus resume and holdout evaluation.
Five built-in schemes — fruit picking, weeds, pests, steel-pipe QC, textile QC — and the same request can go to a local model or TypeSafe's official Jev.
MIT, 57 stars, 8 forks, created 2026-10-05, pushed 2026-10-08; formerly laya-workbench, with API keys stored only on the local machine.

Quickstart

bash

tar -xzf laya-opencv-*.tar.gz
cd laya-opencv
bash install.sh
bash start.sh

Watch out

README reviewed 2026-10-17 and truncated while describing headless-server use, so the full training walkthrough, exact GPU requirements beyond the 4 GB VRAM note, and pricing are not in the digest. The steel-pipe QC screenshot is a synthetic example the README says the program generated, and no accuracy benchmarks for the shipped schemes appear in the excerpt.

Questions about this entry

Short answers drawn from the linked project or page.

From the Jev guides

Original write-ups on what this entry is for.

Jev at home: voice pipelines, Assist routing, and sentinels

Oct 16, 2026

How Home Assistant conversation agents, macOS voice pipelines, and an advisory sentinel put a fast typed decision in the loop — one call per utterance, thresholds in code, and a daily token budget.

Jev Tools & CLIs#home-assistant#voice#automation

Read article

Search visibility, audited by decisions

Oct 16, 2026

Jev-based SEO tools split the crawl from the call: deterministic crawlers count and parse, while typed judgments rank fixes, place internal links, score search results, and track AI-answer mentions — with costs and caveats attached.

Jev Tools & CLIs#seo#search-visibility#audits

Read article

Jev on the Gateways: Vercel, Workers AI, and OpenRouter

Oct 12, 2026

The request shape barely changes across surfaces; the surrounding contract does. When to call the API directly and when a gateway, a Worker binding, or a router entry earns its extra hop.

Jev Tools & CLIs#gateway#vercel#cloudflare

Read article

From the Jev AI model community

Posts from builders shipping with Jev right now.

Follow @typesafeai

Laya OpenCV

Highlights

Quickstart

Questions about this entry

What is Laya OpenCV?

How does it involve Jev?

How do I install it?

What can I process with it?

Can I train my own decision model?

Do I need a GPU to train?

Is it open source, and where do API keys go?

More Jev resources like this

From the Jev guides

From the Jev AI model community