Docs
CommunityJev Guides & Articles
OpenDecision — deepanwadhwa.github.io
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.
- Format
- Docs
- Reading time
- 3 min
- Published by
- Community
- Author
- deepanwadhwa
- Last verified
- Oct 2026
This is one of several Jev listings for the same project. The canonical entry is OpenDecision — deepanwadhwa.
Highlights
- Runs locally, answers typed questions about text, JSON, documents and images, and returns structured values rather than text.
- Primitives: Choice (option and probabilities), Noul (0-1 score), Score (weighted score) and Relation (supports, contradicts, unknown, conflicted).
- Text context is 512 tokens for Choice, Score and Relation and 8,192 tokens for Noul; images resize to 224 x 224 with 64 tokens per caption.
- Interfaces: in-process Python, POST /v1/systemone, POST /v1/documents/decide and a TypeSafe-compatible endpoint for compatible SDK clients.
- Recorded latency examples: 231 ms for two text questions and 155-318 ms for image examples, medians of five local calls on an Apple M3.
Quickstart
bash
pip install OpenDecisionWatch out
Page text fetched 2026-10-17. The docs state measurements were made on CPU with short inputs and default models, and the latency numbers are the authors' own medians of five local HTTP calls on an Apple M3; image results are recorded demos on simulated tasks with rules using a 0.5 yes/no threshold and a 0.6 minimum Choice probability. The digest captured the introduction page only, so the linked get-started, primitives, image and document pages were not verified.
Questions about this entry
Short answers drawn from the linked project or page.
An open-source semantic decision engine for structured state and documents. It answers typed questions about text, JSON, documents and images, runs locally, and returns structured values instead of generated text; the docs site footer names its licence as Apache 2.0.
Four primitives: Choice selects one option and returns option name and probabilities; Noul tests a statement and returns a score from 0 to 1; Score uses an ordered scale and returns a weighted score and probabilities; Relation compares evidence with a statement and its opposite, returning supports, contradicts, unknown or conflicted.
The docs give pip install OpenDecision or uv add OpenDecision, then a get-started guide covering a Python example and the API. An engine instance answers in the same process, for example engine.choice(state=..., instructions=..., criteria={...}) returning billing in the sample.
Text context is 512 tokens for Choice, Score and Relation, or 8,192 tokens for yes/no Noul, including the question text. Images are resized to 224 x 224 pixels with 64 tokens per caption, and the document API splits longer documents into passages; longer inputs and concurrent requests take more memory.
In-process Python, an HTTP endpoint POST /v1/systemone for state plus typed questions, POST /v1/documents/decide for documents, and a TypeSafe-compatible endpoint so a compatible TypeSafe SDK client can talk to a local server. The site also lists a maintainer guide.
Yes - a Doom demo where OpenDecision picks actions for a bot in ViZDoom's Deadly Corridor, with recordings for Skill 5, Skill 1 and Skill 3. The image examples ask questions about photos, such as wildlife alerts, bin collection and fence repair, and combine the answers into actions.
The docs name deepanwadhwa/OpenDecision and state the project is Apache 2.0. No pricing appears anywhere in the captured page, and the engine is described as running locally, so this entry cannot quote any cost; the digest covers only the introduction page.
Reactions & coverage
Posts, threads, and videos about this entry from around the web.
X: 19 open-source Jev projects
X: Jev repositories worth a look, in Japanese
More Jev resources like this
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.
Jev Models & Reimplementations#community#python#local
Community
deepanwadhwa/OpenDecisionPython
deepanwadhwa/OpenDecisionPythonMapika'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 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/DecisPythonA local System One-compatible decision model on Qwen2.5-0.5B-Instruct: 50.7% accuracy and 0.245 ECE on JevBench public with a 13.1 ms median, built for demos rather than production.
Jev Models & Reimplementations#community#open-models#local
Community
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-flowPythonFrom 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.