Jev Rerank Server
A drop-in rerank API server powered by TypeSafe AI's Jev model, compatible with Cohere, Jina, and Voyage rerank protocols for RAG stacks.
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 drop-in rerank API server powered by TypeSafe AI's Jev model, compatible with Cohere, Jina, and Voyage rerank protocols for RAG stacks.
Implements a Jev/System One-compatible decision inference API with a CUDA engine, official SDK drop-in usage, runnable examples, and reproducible benchmarks.
A TypeSafe-compatible System One API running on Raspberry Pi, powered by llama.cpp Q8 and a torch decision head, 4.8x faster than the official SDK with identical decisions.
Model-agnostic Jev-compatible server offering fast decisions via single-token logprobs, works with any OpenAI-compatible endpoint.
A proof-of-concept server implementing TypeSafe's Jev API using local Laya checkpoints, verified against the real SDK.
A decision model replicating TypeSafe Jev's architecture, compatible with the System One API via typesafe-sdk; includes demo, evaluation, and training scripts.
OpenJev Multimodal is an independent, open-source local implementation of a Jev/SystemOne-compatible API using Qwen and llama.cpp/Metal. It provides Noul, Choice, and Score primitives, vision input, a playground, examples, and small-sample benchmarks.
Decis implements TypeSafe's System One API contract, is compatible with Jev's /v1/systemone endpoint, works with the official typesafe-sdk, and can run open decision models such as Laya and kev.
dan is a Python library and server that turns any open-weight causal LM into a System One model with a Jev-compatible API, reading label-token probabilities in a single prefill. It includes calibration and LoRA training tools.
A native Go runtime for Jev/System One decisions using Gemma, supporting Noul, Choice and Score, with benchmarks and a Web playground.
A self-hosted API and web interface for Laya's structured decision models, compatible with the TypeSafe Jev API format.
Jev-compatible API endpoint based on open models (prefill-only), implementing TypeSafe Jev/System One evaluation endpoints with Qwen3.6-35B-A3B on SGLang.