User repository on GitHub
Tool kits for building with System-One models
- Forks
- 0
- Last pushed
- Updated Oct 2026
Repo stats from the GitHub API, cached Oct 2026.
Jev directory
CommunityJev Repos & SDKs
Question Kit
A TypeScript-first toolkit billed as the standard library for System One models: ready-made classify, extract, pickOne, rank and verify methods, a service-agent runtime that runs conversations turn by turn, and a one-message order parser. A Python port is in preview and the project is in alpha.
- Stars
- 4
- Language
- TypeScript
- License
- MIT
- Reading time
- 2 min
- Published by
- Community
- Author
- Engid
- Last verified
- Oct 2026
- Languages
- TypeScriptPython
- Platforms
- API
- Capabilities
- Open source
Highlights
- Ready-made methods: classify, classifyTree, pickOne, extractValue, extractDate, callFunction, verifyRecord, checkClaim, search, rerank and more.
- Service agent: describe a service as data and code runs the procedure while Jev reads the customer, turn by turn.
- Every reading passes a confidence gate before the agent acts; no language model writes anything in the agent.
- Node 20+ or Bun; npm install question-kit @typesafe-ai/sdk; calls need TYPESAFE_API_KEY and a cache keeps re-runs free.
- MIT license; alpha, not taking pull requests yet; methods were developed against thousands of labeled cases.
Quickstart
bash
npm install question-kit @typesafe-ai/sdk
bun install
echo "TYPESAFE_API_KEY=..." > .env
bun run pizzaWatch out
README reviewed 2026-10-17; the project is an alpha whose APIs will change between 0.x releases, it works with Jev only for now, the evals behind its design are private, and the Python port is a preview with only core plumbing and a first set of methods.
Questions about this entry
Short answers drawn from the linked project or page.
question-kit is a TypeScript-first toolkit described as the standard library for System One models, currently working with Jev. It provides ready-made methods for common jobs, a service-agent runtime, an order parser, a cache, a lint for badly formed questions and a client for TypeSafe's API.
You send Jev state and closed questions and it returns a probability per option; question-kit writes the questions and turns answers into typed results with confidence. Methods such as classify, extractValue, pickOne and rerank each follow one of TypeSafe's cookbooks.
Describe a service as data (intents, slots, tools, procedures) and question-kit/service-agent runs the conversation turn by turn: code runs the procedure and Jev reads the customer's intent, values and open judgment calls. It has been through internal evals but has not run on live chats yet.
Install with npm install question-kit @typesafe-ai/sdk, or for the Python preview pip install "question-kit[typesafe]". The pizza shop example runs with bun install, echo TYPESAFE_API_KEY=... > .env and bun run pizza, or bun run pizza --fake with no key.
classify, classifyTree, pickOne, extractValue, extractDate, callFunction, verifyRecord, checkClaim, search, rerank, matchRecords, screen, filterPassages, recoverStructure, rubric, featurize and stability, plus rankByScore and rankPairwise for ordering lists.
It is an alpha whose APIs will change between 0.x releases, works with Jev only for now, and is not taking pull requests yet. The core methods and the agent were measured against thousands of labeled cases, but the evals themselves are private, and the Python port is in preview.
The Engid organisation publishes it under the MIT license; the repository was created 2026-10-06 and pushed 2026-10-17. The README credits Stately's jevspresso demo as the idea behind parsing orders with closed questions, with no code copied.
More Jev resources like this
A Python library that turns Jev-style decision models into ordinary function calls: check returns a bool, classify picks one option, label picks every option that applies, and score rates on a scale, with probabilities behind every answer.
Jev Repos & SDKs#community#github#python
Community
jlowin/vibecheckPython
jlowin/vibecheckPythonA typed TypeScript client for Jev System One models that sends a state plus Choice, Score or Noul questions to POST /v1/systemone and returns typed answers, with presets for TypeSafe, Zen and custom gateways.
Jev Repos & SDKs#community#github#typescript
Community
sc0d3r/jev-systemoneTypeScript
sc0d3r/jev-systemoneTypeScriptThe official JavaScript and TypeScript SDK for the TypeSafe API, at version 0.7.0: typed Choice, Score, Noul, and box questions, experimental image inputs, and zstd request compression on Node.js 22.
Jev Repos & SDKs#community#typescript#sdk
Community
A System One layout optimizer for decision-model workloads: it reorders rows and fields so a prefix-caching backend reuses input computation, then returns per-record decisions in original order. Its bundled backend serves AutoTrust JEV-9B through vLLM.
Jev Repos & SDKs#community#github#python
Community
0814wdwd/solo_jevPython
0814wdwd/solo_jevPythonA Bangla-focused adaptation of Laya, the non-autoregressive typed-decision model: targeted data preparation, fine-tuning and evaluation for classification, routing and scoring on Bangla, Banglish and code-switched text, with the checkpoint still to come.
Jev Repos & SDKs#community#github#python
Community
shamspias/laya-banglaPython
shamspias/laya-banglaPythonHTTP server for the Mica v0.5 4B Preview decision model: GGUF weights run through llama.cpp and answer noul, choice and score questions with calibrated probabilities over the TypeSafe /v1/systemone format.
Jev Repos & SDKs#community#github#python
Community
akivet/MicaPython
akivet/MicaPythonFrom 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.