Jev directory
CommunityJev Models & Reimplementations
Decidealot on Docker Hub
psyb0t's Docker image that serves local Laya, Von, a CLM projection head, or hosted Jev through one TypeSafe-compatible HTTP and MCP endpoint, returning typed choice, score, and noul answers with model probabilities.
- Reading time
- 2 min
- Published by
- Community
- Author
- psyb0t
- Last verified
- Oct 2026
Highlights
- Serves TypeSafe-compatible POST /v1/systemone and /v1/systemone/batch, and MCP Streamable HTTP with system_one, system_one_batch, list_models, unload_models.
- At startup it downloads and verifies every enabled local bundle, then loads only the provider a request selects and unloads it after the configured idle period.
- Defaults enable Laya and Von; CLM adds a small local projection head over one configured Qwen3-8B embeddings endpoint, with temperature defaulting to 1.
- Hosted Jev enables automatically when DECIDEALOT_TYPESAFE_API_KEY is set, caches the TypeSafe catalog for 60 seconds, and needs no release for new model names.
- Quick start: the CPU image runs on loopback with a read-only root filesystem, all capabilities dropped, no-new-privileges, and a /health check.
Watch out
The page states no licence for the container itself, so redistribution terms are not established, while bundled models carry their own licences. Hosted Jev calls are billed by TypeSafe, decisions remain advisory because Decidealot never performs the action, and the project belongs to a single Docker Hub user rather than a vendor.
Questions about this entry
Short answers drawn from the linked project or page.
A Docker image by psyb0t that makes the call for typed classification and scoring: run local Laya, Von, or a CLM projection head, or call hosted Jev, through one TypeSafe-compatible HTTP and MCP endpoint.
Run the documented docker run command, which starts the CPU image on loopback and bind-mounts a single model directory; wait for GET /health before sending decisions, because the first startup downloads and verifies the enabled pinned bundles.
Laya and Von are enabled by default, CLM is available as a projection head over a configured Qwen3-8B embeddings endpoint, and hosted Jev enables automatically when DECIDEALOT_TYPESAFE_API_KEY is present. GET /v1/models lists the live catalog.
The same container serves MCP Streamable HTTP at /mcp with system_one, system_one_batch, list_models, and unload_models tools. Point an MCP client at the URL, adding an Authorization bearer header unless DECIDEALOT_API_KEY is empty.
No. It returns typed answers with probabilities and does not perform the action; the readme says your caller checks the chosen label and applies its own threshold, for example allowing an action only when choice is allow and probabilities.allow is at least 0.95.
CLM config accepts a temperature defaulting to 1, with finite values above 0 and at most 100; lower values sharpen its probability distribution but do not make the answer more correct. Laya, Von, and hosted Jev accept only omitted or empty config.
The quick-start command uses a read-only root filesystem, --cap-drop ALL, no-new-privileges, pids and memory limits, noexec tmpfs mounts, and loopback-only publishing; MCP keeps DNS-rebinding protection and needs allowed hosts and origins configured behind a proxy.
Reactions & coverage
Posts, threads, and videos about this entry from around the web.
X: Standard One SH (StandardThinking)
LinkedIn: What multimodal decision models actually change, and what they do not.
LinkedInJevDirectory.org
Hacker News: Decision Index 0.2.1 scores open reproductions of Jev on a frozen suite.
jackbravo4 pointsDecision Index 0.2.1 – 70 open reproductions of TypeSafe Jev's Decision ModelRead the thread on Hacker NewsX: Jev in front of the HeyGen MCP
More Jev resources like this
A local daemon that pulls open decision models and serves them on TypeSafe's /v1/systemone, so a Jev client switches hosts with TYPESAFE_BASE_URL.
Jev Models & Reimplementations#community#rust#open-models
Community
ollaya-dev/ollayaRust
ollaya-dev/ollayaRustA LoRA plus 24-slot decision head over Qwen3.8-27B from Ming Xu's train-jev project: it serves TypeSafe's System One API with 0.851 Kev transfer-v4 accuracy and ECE 0.053, and needs an 80 GB GPU.
Jev Models & Reimplementations#community#open-models#local
Community
A 24 GB-GPU alternative from the train-jev project: Qwen3.5-4B plus a rank-16 LoRA and 24-slot decision head scoring 0.647 PK accuracy, 40.4 Decision Index and 43 ms p50 latency, ahead of Kev-4B at the same size.
Jev Models & Reimplementations#community#open-models#local
Community
Singularity Compute's fine-tune of Google's diffusiongemma-26B-A4B-it, a discrete-diffusion mixture-of-experts, served from an 18 GB NVFP4 checkpoint that answers choice, noul and score questions with calibrated distributions on one 48 GB GPU.
Jev Models & Reimplementations#community#open-models#local
Community
StartLux Labs' 27B AutoThink decision model reads a state (text, JSON or up to 8 images) and typed questions, answers in one pass and only writes a thought when the top probability falls below the 0.6 gate; self-reported Decision Index 0.3 public score 73.37.
Jev Models & Reimplementations#community#open-models#local
Community
Mapika's NVFP4 fine-tune of Gemma-4-31B-it answers typed Choice, Noul and Score questions with calibrated probabilities and no generated text, self-reporting a 64.27 chance-corrected public Decision Index and 39 ms median latency on a B300.
Jev Models & Reimplementations#community#open-models#local
Community
From 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.