Docs
CommunityJev Integrations & Adapters
Jev on LiteLLM Decision Routes
LiteLLM serves decision models on two routes that share one provider list: /v1/decisions for OpenAI-format bodies and /v1/systemone for System One bodies, with Jev available as the typesafe/jev-latest provider.
- Format
- Docs
- Reading time
- 2 min
- Published by
- Community
- Author
- LiteLLM
- Last verified
- Oct 2026
- Platforms
- API
- Capabilities
- Self-hostedRequires API key
Highlights
- Two routes share one provider list: /v1/decisions for OpenAI-format bodies and /v1/systemone for System One bodies, translated both ways.
- Jev is provider typesafe/jev-latest with TYPESAFE_API_KEY, and sending the sample questions to model jev returns the same response shape.
- Cost tracking via x-litellm-response-cost, logging, virtual keys, budgets, rate limits, load balancing, and fallbacks all work; streaming does not.
- Available from v1.104.2 on the 1.104.x line and v1.105.0-rc.3 and later; on v1.106.0-dev.1 send System One bodies to /v1/systemone.
- Self-hosted Laya and Bespoke Nimble are priced at $0 because you pay for the hardware, so set Custom Pricing to track spend.
Watch out
Vendor docs for a self-hosted proxy: the page states no licence, hosted-provider pricing, or benchmarks. Image input on TypeSafe needs the TypeSafe-Danger-Zone-Images: true header, streaming is unsupported, and TypeSafe images, box questions, and the danger-zone headers were unreleased at capture.
Questions about this entry
Short answers drawn from the linked project or page.
POST /v1/decisions speaks OpenAI's Decisions format with predicate, choice, score, and box questions, and POST /v1/systemone speaks System One with noul, choice, score, and box questions. They share one provider list, and LiteLLM translates between the formats.
Add a model_list entry with model typesafe/jev-latest and api_key os.environ/TYPESAFE_API_KEY, then run litellm --config config.yaml, which starts the proxy on http://0.0.0.0:4000. A TYPESAFE_API_BASE can override the default endpoint.
OpenAI, TypeSafe, Fireworks AI in private preview, Perplexity, Cloudflare Clef, Databricks with its serving endpoints or ai_decide function, Microsoft Foundry, plus self-hosted Strands Decider, vLLM, Laya, and Bespoke Nimble.
No. Streaming is listed as unsupported for decision models; answers come back in one response with type-specific fields such as probability, choice, score, per-option distributions, confidence, and usage token counts.
Cloudflare model names without an @cf/ prefix expand to @cf/cloudflare/<model>, and the result envelope is unwrapped. On Databricks the model is the serving endpoint name that receives /invocations, while databricks/ai_decide calls the AI Function and reports no token usage.
v1.104.2 and later on the 1.104.x line, and v1.105.0-rc.3 and later. On v1.106.0-dev.1, which served the System One format at /v1/decisions, send System One bodies to /v1/systemone; Databricks and Microsoft Foundry land in the next build.
Yes. Decision routes use normal proxy key auth and support virtual keys, budgets, rate limits, logging, cost tracking returned in the x-litellm-response-cost header, and router load balancing and fallbacks like /chat/completions.
Reactions & coverage
Posts, threads, and videos about this entry from around the web.
X: Jev 1.14 is now on OpenRouter
X: Jev lands on the Vercel AI Gateway
X: Jev as an LLM guardrail inside Agentgateway
YouTube: Greg Isenberg and Ryan Vogel on the businesses Jev unlocks, and a path in through the Vercel AI Gateway.
More Jev resources like this
Akka's post walks from a Jev triage example - eight questions with probabilities - to an effects API judgment call and an AI Gateway that can route to hosted Jev or a self-hosted OpenJev model.
Jev Integrations & AdaptersArticle#community#java#akka
Community
Concentrate's beta /v1/decisions endpoint serves TypeSafe's Jev with the same request and response shapes as TypeSafe's System One API and OpenRouter, returning calibrated probabilities for noul, choice, and score questions.
Jev Integrations & AdaptersDocs#community#gateway#api
Community
Fly.io's TypeSafeAI connector gives Sprites access to Jev without the API key entering the machine: the Sprites gateway attaches the organization's encrypted key, and each Sprite sets only TYPESAFE_BASE_URL plus a placeholder key.
Jev Integrations & AdaptersArticle#community#fly-io#agents
Community
OpenCode's Zen gateway serves Jev 1.13 on its own /v1/systemone endpoint with an OpenCode key: jev-1.13 paid and jev-1.13-free for a limited time, documented with noul, choice, and score examples.
Jev Integrations & AdaptersDocs#community#opencode#gateway
Community
Pydantic's AI Gateway forwards Jev natively as a bring-your-own-key TypeSafe provider, so one gateway key covers spend views, limits, and input guardrails while TypeSafeModel turns each output field into a question.
Jev Integrations & AdaptersArticle#community#pydantic#gateway
Community
A session-aware, OpenAI-compatible gateway that routes chat completions across provider pools by cost, quality, and capability rules, optionally asks configured decision providers typed choice questions with a deterministic fallback, and records requests, decisions, and outcomes in SQLite.
Jev Integrations & Adapters#community#catalog-import#logicrw
Community
texasoct/jev-gatewayTypeScript
texasoct/jev-gatewayTypeScriptFrom 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.