Jev directory
CommunityJev Models & Reimplementations
Koin
Satirical Python script that answers a query with DO or DO_NOT plus a random confidence percentage, presented as a natural alternative to Jev. MIT-licensed, Python 3.8+, one dependency, no API keys or model weights.
- Stars
- 0
- Reading time
- 2 min
- Published by
- Community
- Added
- Nov 2026
Highlights
- Returns DO or DO_NOT with a confidence percentage produced by a single random.random() call
- Advertises 50% accuracy and sub-millisecond inference with no model weights to load and no GPU
- Requires Python 3.8+ and one dependency, rich; no API keys and no configuration
- Interactive CLI loop prints a result block and stops when you type /exit
- Repository states plainly that Koin is satire and not affiliated with TypeSafe AI
Questions about this entry
Short answers drawn from the linked project or page.
Koin is a small Python program that takes a query and returns either DO or DO_NOT together with a confidence percentage. Its own README describes it as satire: a coin toss wearing a lab coat, marketed as a natural alternative to Jev.
Clone the repository, change into the Koin directory, install the single dependency with pip install rich, then run python3 koin.py. The script opens an interactive prompt where you type a question and read the printed result block.
No. The digest states there are no model weights to load, no GPU to rent, and no API keys. The decision layer is one random.random() call, and the tokenizer output is ignored entirely.
The repository badges and benchmarks section state 50% accuracy. The README frames that number as the feature rather than a defect, since the output is a coin toss rather than a learned decision.
No. The README states that Koin is satire and is not affiliated with, endorsed by, or remotely comparable to TypeSafe AI. It is a joke project that references Jev as a foil.
The repository carries an MIT license badge, and a LICENSE file is present in the project. The README also links a CONTRIBUTING.md and welcomes pull requests.
The README explicitly warns against routing production traffic, medical decisions, or money through it. It is a satirical utility, not a calibrated decision model, so it should not back any real decision path.
Reactions & coverage
Posts, threads, and videos about this entry from around the web.
X: Standard One SH (StandardThinking)
LinkedIn: What multimodal decision models actually change, and what they do not.
LinkedInJevDirectory.org
Hacker News: Decision Index 0.2.1 scores open reproductions of Jev on a frozen suite.
jackbravo4 pointsDecision Index 0.2.1 – 70 open reproductions of TypeSafe Jev's Decision ModelRead the thread on Hacker NewsHacker News: Pointer to a DiffusionGemma plus vLLM tutorial for a Jev-shaped classifier.
1 pointHow to build a Jev-style classifier with DiffusionGemma and vLLMRead the thread on Hacker NewsMore Jev resources like this
Open Jev-style decision model built on LiquidAI's LFM2.5-350M backbone, a clone of kev. Serves the System One API contract so the TypeSafe SDK works against a local server, returning probabilities for typed questions from one forward pass.
Jev Models & Reimplementations#open-models#python#local
Communityfranckverrot/lev
Russian-first open typed-decision model on Hugging Face. Given a context and a set of candidate options it returns a calibrated probability and picks one, covering yes/no, choice, and ordinal scale decisions. Built on mmBERT-base with 307M parameters under an MIT license.
Jev Models & Reimplementations#open-models#calibration#classification
Community
A 561M-parameter XLM-RoBERTa-large decision model given one extra training epoch on Jev's answers, with about 23% of examples black-box distilled from Jev's public API. Supports noul, choice, and score questions in 18 languages and loads through the bundled naluri.py.
Jev Models & Reimplementations#open-models#calibration#benchmark
Community
An independent, from-scratch reconstruction of a Jev-style decision model that never generates text: a state plus typed questions returns a probability for every option. Trained on Qwen3-0.6B and 1.7B backbones with LoRA and served through a system_one-shaped API.
Jev Models & Reimplementations#open-models#python#research
CommunityES7/JEV-from-Scratch
A from-scratch Jev-style decision model on a frozen Qwen2.5-3B backbone with a custom decision head and hand-rolled LoRA adapters. It returns typed action distributions for routing, safety gating, and RAG filtering, reporting 83.38% accuracy at 23.95 ms p50 latency.
Jev Models & Reimplementations#open-models#python#routing
Communityabirshhishirroy/mini-jev
A native MLX runtime for Laya's open typed-decision weights: short decisions on Apple Silicon, with no generated tokens, PyTorch, or cloud API.
Jev Models & Reimplementations#community#python#open-models
Community
mizorewww/laya-mlxPython
mizorewww/laya-mlxPythonFrom the Jev guides
Original write-ups on what this entry is for.
From the Jev AI model community
Posts from builders shipping with Jev right now.