vllm2jev
A lightweight script that adapts vLLM services to a Jev-style interface, enabling Choice, Score, and Noul decisions via logprobs.
OPEN SOURCE · CURATED REPOSITORIES
Curated Jev SDKs, tools, demos, integrations, benchmarks, and research projects. Repository metrics come from GitHub; titles, summaries, and categories are editorial.
Ranked by GitHub Stars
Ranked by recent Star growth and update activity
A lightweight script that adapts vLLM services to a Jev-style interface, enabling Choice, Score, and Noul decisions via logprobs.
Uses TypeSafe AI's Jev (System One) model to drive the ego lite browser: one indexed element table in, one operation and target out per request. Includes comparative benchmarks showing ~2x faster than per-step LLM loops.
A 3D digital human demo powered by the TypeSafe Jev decision model, featuring a dual-system architecture with natural language control and tool invocation.
Open-source Python framework integrating TypeSafe AI's Jev/System One probabilistic engine with LLM dual pipelines for CSV field type recognition and data cleaning; tests show Jev and LLM classification conclusions align.
A-share technical analysis workspace powered by TypeSafe Jev/System One, integrating AKShare/Tushare for traceable Buy/Hold/Sell assessments.
An evidence-graded deep dive into TypeSafe AI's Jev model, covering interface, verification, reproduction comparisons, and a field guide, with runnable probe tools; bilingual EN/中文.
TypeSafe Jev integration providing a core decision library, DeepSeek Harness plugin, and MCP server; offline by default, with tests and real API verification.
An integration example of Jev autonomously playing Stardew Valley, including a dsh plugin, CLI, SMAPI Mod, and a runnable mock farm.
A 2048 web lab where every move is driven by TypeSafe AI's Jev decision model, showing probability, confidence, latency, and cost, with no heuristic fallback.
jev-assist is a coding-agent skill that uses TypeSafe Jev/System One to rank repository files by task relevance, detect convention drift, and flag risky diffs before commit. It includes setup, validation, and pre-commit hook wiring.
Real-world benchmark of TypeSafe's Jev for robot fleets and edge hardware, with 300 measured API calls, cost analysis, and reproducible scripts.
A community-maintained curated list of Jev / TypeSafe System One projects, SDKs, tutorials, evaluations, and Chinese resources, with a bilingual catalog and usage guides.