User repository on GitHub
Self-host Kev-4B, the open-weight SystemOne decision model, behind a TypeSafe-compatible POST /v1/systemone API via docker compose
- Forks
- 0
- Last pushed
- Updated Sep 2026
Repo stats from the GitHub API, cached Oct 2026.
Jev directory
CommunityJev Repos & SDKs
Kev POC
A docker compose recipe that self-hosts Kev-4B, the open-weight SystemOne decision model, behind a TypeSafe-compatible POST /v1/systemone API, downloading the adapter and Qwen3.5-4B-Base weights on first start and caching them in a named volume.
- Stars
- 0
- Language
- Dockerfile
- License
- APACHE-2.0
- Reading time
- 2 min
- Published by
- Community
- Author
- dashaun
- Last verified
- Oct 2026
Works forJev for Classification
Used inJev for Software & DevTools
- Platforms
- API
- Capabilities
- Open sourceSelf-hosted
Highlights
- Self-hosts Kev-4B, the open-weight SystemOne decision model, behind a TypeSafe-compatible POST /v1/systemone API.
- Starts with one command: docker compose up --build -d, then check it with curl -s localhost:8008/v1/models.
- First start downloads the adapter and Qwen3.5-4B-Base weights into the kev-hf volume using the maker's kev.serve.
- Binds 0.0.0.0:8008 so other machines can call it; the README covers base_url, WSL2 IP changes and firewall rules.
- With no KEV_API_KEY set, anyone who can reach port 8008 can call the API.
Quickstart
bash
docker compose up --build -d
curl -s localhost:8008/v1/modelsWatch out
README and GitHub metadata reviewed 2026-10-17; the captured excerpt cuts off inside the security note and the Configuration section is not shown, so environment options, hardware requirements and performance claims are unavailable.
Questions about this entry
Short answers drawn from the linked project or page.
A docker compose recipe that self-hosts Kev-4B, the open-weight SystemOne decision model, behind a TypeSafe-compatible POST /v1/systemone API, so existing Jev clients can point at your own server.
The open-weight SystemOne decision model published as jared-palmer/kev; this POC runs the maker's own kev.serve server and downloads the adapter plus Qwen3.5-4B-Base weights on first start, caching them in the kev-hf volume.
One command from the repository: docker compose up --build -d. The container uses CUDA torch and flash-linear-attention when available, and the first start also pulls the model weights.
The README shows curl -s localhost:8008/v1/models, and the same host and port accept POST /v1/systemone calls with a JSON body; a curl example for posting is also included.
Yes. The container binds 0.0.0.0:8008, so any host that can reach the machine can call it; find the IP with hostname -I and point the TypeSafe SDK base_url at it. WSL2 IPs change on reboot and Windows Firewall may need inbound TCP 8008.
No. With no KEV_API_KEY set, the API is open to anyone who can reach port 8008; the README says to set the key before exposing the container, and the captured sentence is cut off right there.
The captured README stops inside the security note, so the Configuration section, GPU requirements, environment variables and any performance numbers are not visible; the repo has 0 stars and was created and pushed 2026-09-30.
More Jev resources like this
A self-hosted System One API billed as drop-in compatible with TypeSafe Jev: a bidirectional encoder with direct decision heads returns calibrated probabilities and structured choices in a single forward pass of about 33 ms.
Jev Repos & SDKs#community#github#fastapi
Community
ZulfiFazhar/system-oneHTML
ZulfiFazhar/system-oneHTMLA 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 Laravel-ready PHP SDK for Laya, a self-hosted multilingual decision engine described as a Jev AI alternative: classify text in 100+ languages and get typed enums, ints and bools back without paying per-token fees.
Jev Repos & SDKs#community#github#php
Community
marcreichel/laya-phpPHP
marcreichel/laya-phpPHPHTTP 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/MicaPythonSelf-hosted Python SDK and System One-compatible API for runtime-defined Choice, Boolean and Score decisions on pinned open models, across Torch, vLLM, MLX, llama.cpp and WebGPU. Each result records probabilities, token usage, timing, model revision and prompt version, with committed row-level benchmarks.
Jev Repos & SDKs#community#sdk#self-hosted
Community
chengyongru/fastjevPython
chengyongru/fastjevPythonServing code for the Jeff 1.0 family of Jev-style decision models: a state, a typed question and its options go in, and one forward pass returns a probability for each option. Four open models from 4B to 31B parameters are served the same way.
Jev Repos & SDKs#community#github#python
Community
j-geuter/jeff-servePython
j-geuter/jeff-servePythonFrom 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.