User repository on GitHub
OpenDecision is an open-source semantic decision engine like typesafe's jev.
- Forks
- 3
- Last pushed
- Updated Oct 2026
Repo stats from the GitHub API, cached Oct 2026.
Jev directory
CommunityJev Models & Reimplementations
OpenDecision · deepanwadhwa
An Apache-2.0 decision engine with a TypeSafe-compatible endpoint, document answers with source passages, and a ViZDoom demo that picks actions from structured game state.
- Stars
- 101
- Language
- Python
- License
- APACHE-2.0
- Reading time
- 3 min
- Published by
- Community
- Author
- deepanwadhwa
- Last verified
- Sep 2026
About this project
Open-source semantic decisions for structured state and documents. Documentation · Doom demo · Get started · API · Examples. OpenDecision is the open-source equivalent of TypeSafe's Jev. Send it application state or a document and typed questions. It returns structured answers that code can use directly. Jev and OpenDecision use the same core pattern: state plus typed questions in, structured answers out. Both provide Choice, Noul, and Score. Choice to select one option from a list.
Quoted from the project's README. Read the full README.
- Languages
- Python
- Capabilities
- Open sourceSelf-hostedRuns locally
Highlights
- Question types are Choice, Score, Noul, and Relation: supports, contradicts, unknown, or conflicted.
- A TypeSafe SDK compatible endpoint is part of the API surface, alongside a Python API and a FastAPI server.
- The Doom demo reads health, ammo, kills, positions, and recent damage. The recordings are fixed-seed, and no win rate is measured.
- The document API splits long documents and returns the passages it used. A GDPR retrieval set answers 9 of 10 objective questions.
Quickstart
bash
opendecision serveWatch out
Apache-2.0. The README calls it a developer preview and says to treat scores as uncalibrated until you fit thresholds on your own data. Not affiliated with TypeSafe.
Questions about this entry
Short answers drawn from the linked project or page.
Same request pattern, different weights. The README calls it the open-source equivalent of Jev and offers the same Choice, Noul and Score primitives plus a Relation type returning supports, contradicts, unknown or conflicted. The default backend is ModernBERT-large-zeroshot-v2.0, an NLI classifier, so answers come from classification scores rather than a chat model.
Calibrate them yourself. Version 0.1.2 is called a developer preview and the README says to treat the model scores as uncalibrated, evaluating the model and thresholds on your own data before using results in an automated decision process. It also needs Python 3.13 or later, and the model downloads from Hugging Face on first use.
Yes, that is what the document API is for. It splits long documents, finds relevant passages, answers each question, and returns the passages used. On the bundled GDPR retrieval experiment a 54,171-character document split into 31 sections answered 9 of 10 objective questions correctly. Whether that holds on your documents is not documented.
Just recordings, and the README is explicit. The clips are fixed-seed recordings at several skill levels and no win rate has been measured. The input is structured game state such as health, ammo, kills, positions and recent damage, and a disclosed tactical router plus a target tracker does much of the work, so this is not Jev driving an agent unaided.
Choice to select one option, Noul to measure whether a statement is true, Score to place state on an ordered scale, and Relation to report supports, contradicts, unknown or conflicted. It also has a document API with evidence retrieval and source passages.
Python 3.13 or later. pip install OpenDecision, then opendecision serve to start the FastAPI server at 127.0.0.1:8000. The default backend is MoritzLaurer/ModernBERT-large-zeroshot-v2.0, which downloads from Hugging Face on first use.
POST /v1/documents/decide has one noul_mode: binary always chooses true or false, three_way returns supports, contradicts or unknown, and both (the default) runs both and reports confirmed, tentative or conflicted, with a conflict returning answer null.
Reactions & coverage
Posts, threads, and videos about this entry from around the web.
Hacker News: Show HN: OpenDecision, a local decision model with a Doom demo.
dwa35926 pointsShow HN: OpenDecision – a 400M zero-shot model makes local decisions, plays DoomRead the thread on Hacker NewsHacker News: Skeptical thread on launch hype, prior art, and open Jev-like stacks such as OpenDecision.
nandakishor_ml1362 pointsI built non-autoregressive decision models with RL a year agoRead the thread on Hacker NewsX: Standard One SH (StandardThinking)
LinkedIn: What multimodal decision models actually change, and what they do not.
LinkedInJevDirectory.org
Other listings of this project
Forks and copies of this project in the directory. This entry is the canonical one.
The documentation site for OpenDecision, an open-source decision engine that runs locally and answers typed questions about text, JSON, documents and images, exposing Python, HTTP and TypeSafe-compatible interfaces with Choice, Noul, Score and Relation primitives.
Jev Guides & ArticlesDocs#community#docs#local
Community
More Jev resources like this
A local yes/no decision model that speaks Jev's wire format on a laptop CPU, refusing Choice and Score until those question types ship.
Jev Models & Reimplementations#community#python#open-models
Community
feder-cr/jevC++
feder-cr/jevC++A local server that reads next-token probabilities from a fine-tuned Spark model and returns Choice, Score, and Noul answers with zero generated tokens.
Jev Models & Reimplementations#community#python#open-models
Community
Rizzo-AI-Academy/rizzo-flowPython
Rizzo-AI-Academy/rizzo-flowPythonA self-hosted System One server: one /v1/systemone contract, a Docker image each for Laya and kev, and a playground where snake, dino, and tetris call that endpoint.
Jev Models & Reimplementations#community#python#local
Community
chaitin/DecisPython
chaitin/DecisPythonMapika's vision decider: decider-2b v5 text weights transplanted into the Qwen3.5-2B vision-language model, tuned so an image plus a lettered question returns a calibrated probability at one answer slot with no generation. Breakout from pixels scores 41, above its teacher's 22.
Jev Models & Reimplementations#community#open-models#local
Community
A local /v1/systemone server that reads label probabilities from a GGUF chat model, with prompt repetition and a confidence shrink toward uniform.
Jev Models & Reimplementations#community#python#open-models
Community
Mushroom-Systems/lichenPython
Mushroom-Systems/lichenPythonA Python server in front of llama-server that reads one-token label probabilities and exposes them as POST /v1/systemone.
Jev Models & Reimplementations#community#python#open-models
Community
khimaros/verdictPython
khimaros/verdictPythonFrom 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.