Docs
CommunityJev Integrations & Adapters
JEV-1.13.0 on B.AI
B.AI's docs page for JEV-1.13.0: a B.AI-hosted route to TypeSafe's System One model at $0.042 per million input tokens, with Choice, Score, and Noul questions and a 64K-token request budget.
- Format
- Docs
- Reading time
- 2 min
- Published by
- Community
- Author
- B.AI
- Last verified
- Oct 2026
- Platforms
- API
- Capabilities
- Requires API key
Highlights
- The interface is POST https://api.b.ai/v1/decisions with state, model, and questions; jev-latest currently resolves to jev-1.13.0.
- Choice supports up to 255 options, Score takes 2 to 10 ordered levels, and Noul returns 0 to 1 with no separate confidence field.
- Context limits are 64K tokens for the state plus all questions, and 32K for the state plus the longest individual question.
- Pricing is $0.042 per million input tokens, settled at 1 USD = 1,000,000 credits; Jev is billed only for input tokens.
- Recommended uses include routing and triage, RAG passage filtering, content screening, and tool selection over known candidates.
Watch out
B.AI's provider documentation; prices are called standard reference prices that may be lower through top-up bonuses, and final billing follows the platform display. English is the primary language, other languages including CJK need workload-specific validation, and the page states no licence or rate limits.
Questions about this entry
Short answers drawn from the linked project or page.
TypeSafe AI's System One model served through B.AI: it evaluates supplied context against typed questions and returns choices, scores, and probabilities rather than prose. jev-latest currently resolves to the same model as jev-1.13.0.
POST https://api.b.ai/v1/decisions with three fields: model, state as a string, JSON object, or array, and questions, each mapping a question id to a Noul, Choice, or Score definition. The Decisions API reference has the details.
Choice supports up to 255 options, Score takes 2 to 10 ordered levels, and Noul returns a probability from 0 to 1. Choice and Score include a probability distribution and derived confidence; Noul has no separate confidence field.
At $0.042 per million input tokens, with 1 USD settling to 1,000,000 credits deducted from the account balance. Jev is billed only for input tokens, and the cache-write, cache-read, output-token, and web-search billing fields do not apply.
Request routing and support triage, RAG passage filtering to assess relevance and contradictions before content reaches a generative model, content screening against policy criteria, and tool selection over known functions and argument candidates.
Arithmetic, counting, date comparisons, and multi-hop reasoning are called unreliable; keep exact computation in code, use narrowly scoped criteria, and remember schema conformity does not guarantee a correct judgment.
English is the primary training language, and other languages, including CJK text, require workload-specific validation. The page lists no licence, no rate limits, and no data-handling terms for the model.
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
LiteLLM's proxy adds TypeSafe guardrail hooks and a Jev complexity classifier for policy judgments and request routing across 100+ model providers.
Jev Integrations & Adapters#community#python#gateway
Community
BerriAI/litellmPython
BerriAI/litellmPythonLiteLLM 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.
Jev Integrations & AdaptersDocs#community#litellm#gateway
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
A Magpie plugin that adds live free-model catalogs for OpenCode Zen and Kilo Gateway as two non-conflicting providers, with native Jev System One support. Model lists come from each gateway's live /models endpoint, so changes need no plugin update.
Jev Integrations & Adapters#community#github#gateway
An MCP gateway that routes tool selection through Jev: for each request, Jev answers one yes/no question per server and per tool, and the LLM only sees the tools that passed. The repo's benchmark covers 649 tools at 1/32 of a comparable routing cost.
Jev Integrations & Adapters#community#github#mcp
Community
ini8labs/jev-mcp-routerPython
ini8labs/jev-mcp-routerPythonLocal tool-routing classifier for coding agents: a gateway and MCP tool that ask Jev which tool to call next, keep your agent's own model and login, and log every decision so you can start by observing.
Jev Integrations & Adapters#community#catalog-import#logicrw
Community
felpsdev/jev-classifierTypeScript
felpsdev/jev-classifierTypeScriptFrom 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.