questions: { severity: score, needs_tool: boolean, }
JevRouter
A lightweight Jev-powered router for models, tools, and subagents

Every project here really calls Jev in its source · refreshed daily
questions: { severity: score, needs_tool: boolean, }
A lightweight Jev-powered router for models, tools, and subagents
questions: { destructive: boolean, exfiltration: boolean, beyond_scope: boolean, // +1 }
TypeSafe Jev as a decision layer for the Pi coding agent: a measured tool-call gate plus jev_ask for typed, calibrated answers
questions: { stale: choice, }
Bounded TypeSafe Jev workflows for coding agents.
questions: { relevant: boolean, }
Classify first. Read selectively. A portable agent plugin and MCP tool for batch text classification.
questions: { intent: choice, reuse_cache: boolean, needs_subagent: boolean, // +2 }
Connect TypeSafe Jev to Grok Bot as a cheap decision layer - usage gates, skill template, examples
questions: { refund: boolean, team: choice, }
ACP and MCP adapter that bridges TypeSafe Jev with any LLM — computer use and typed decisions alongside Codex, Claude, Grok, and OpenCode.
questions: { ok: boolean, }
MCP server that puts TypeSafe Jev on the coding loop in Cursor, Codex, and any MCP client
questions: { is_billing: boolean, category: choice, }
Semantic tool routing and typed System One decisions for the Pi coding agent using TypeSafe Jev
questions: { directed: boolean, kind: choice, }
Auto mode for every coding agent, built on Jev: risk-scores every tool call with session context (deny / ask / allow), flags prompt injection in results, checks skills and plugins. Claude Code, Codex, Copilot, Gemini, Cursor, pi, OpenCode, ACP.
questions: { is_urgent: boolean, department: choice, frustration: score, }
Jev decision layer for agents: MCP server, embeddable DecisionModel library, and an escalate-only Claude Code plugin (TypeSafe AI's Jev)
questions: { skill: choice, }
Jev-assisted file retrieval and request caching for faster Pi workflows
questions: { claims_done: boolean, claims_verified: boolean, verification_applies: boolean, // +1 }
Claude Code Stop hook that blocks an unverified done: reads the transcript for evidence, asks Jev once, fails open on everything else
questions: { intent: boolean, }
TypeSafe Jev as the pi coding agent's quiet decision layer
questions: { is_bug: boolean, team: choice, urgency: score, }
Claude Code / Codex / pi plugin that hands agent steps needing no text output to Jev (TypeSafe's judgment model) — measured p50 ~230 ms and ~$0.02 per 1,000 judgments, with typed escalation back to the LLM
questions: { intent: choice, isUrgent: boolean, }
Flue agent routing with TypeSafe Jev through Cloudflare AI Gateway
questions: { context_needed: choice, }
Pi coding-agent extension: TypeSafe Jev checks for tool calls, tool outputs and replies (prompt injection, approvals, secret scrubbing, task pinning)
questions: { q1: boolean, }
A small, extensible decision-to-action harness for TypeSafe Jev
questions: { connected: boolean, }
Jev DSH 决策引擎|面向 Agent Harness 的结构化决策插件。原生支持 DeepSeek Harness,通过 iPolloWork 支持 OpenCode、Codex Harness。
questions: { is_impossible: boolean, }
Jev-powered decision layer for DeepSeek Harness
questions: { is_urgent: boolean, frustration: score, }
Claude Code plugin that scores review findings, debug hypotheses and design options with TypeSafe's Jev — calibrated probabilities instead of one more opinion.
questions: { is_urgent: boolean, }
Agent-facing TypeSafe Jev (System One) for the DSH Web GUI. The model itself
questions: { needs_skill: boolean, }
Claude Code mod that routes decisions to TypeSafe's Jev model: ranks installed skills per prompt, and answers the agent's own this-or-that questions when confident.
questions: { category: choice, reached_assertion: boolean, missing_context: boolean, }
Open-source Codex plugin for TypeSafe Jev decision consultation, failure diagnosis, and evidence-based completion review
questions: { complexity: score, needsPlanning: boolean, }
Throwaway Jev demo: route coding tasks to Grok Build or Codex Astra
questions: { destructive: boolean, exfiltration: boolean, privilege: boolean, // +4 }
Open auto mode for AI agents — a calibrated tool-call firewall powered by TypeSafe Jev. Ships as a Claude Code hook
questions: { category: choice, bug_severity: score, has_repro_steps: boolean, // +2 }
MCP server and agent skill for the TypeSafe AI System One API (Jev): decompose a judgment into Choice / Score / Noul questions, lint them, measure on labelled data, and put calibrated thresholds in code
questions: { relevant: boolean, category: choice, }
PiJev: a terminal coding agent with Jev in the loop — Jev ranks the repository's files before the first call, picks skills and triages failures; your coding model writes the code. Built on Pi.
questions: { ping: boolean, }
Yunus Pi is a heavily customized harness experience built on top of a forked Pi harness (pi.dev). It uses patches, extensions, skills, and custom provider configs to customize the overall experience. It uses advanced machine learning, Jev, Cactus Needle 3 and similar technologies.
questions: { action: choice, asset: choice, condition_operator: choice, }
Reference prototype exploring agentic commerce: TypeSafe/Jev → Axiom → Argent → Silverscript on Kaspa.
questions: { severity: score, area: choice, }
An eval-first MCP server for TypeSafe's Jev, a System One model that returns typed judgments (noul, choice, score) with probabilities instead of generated text.
questions: { destructive: boolean, secrets: boolean, }
Jev-powered Judgment layer for Claude Code. Stops paying reasoning prices for if-statements: a PreToolUse hook scores every tool call against a YAML policy you own — allow / deny in ~100 ms, no LLM in the loop. TypeSafe Jev now, local models next. Dry-run by default, calibration table published.
questions: { is_exquisite_and_dynamic: boolean, animation_verdict: choice, primary_deficiency: choice, }
Real-time quality gate and Art Director Warden for Claude Code powered by TypeSafe Jev 1.13 non-autoregressive decision model
questions: { needs_rewrite: boolean, visual: choice, }
Pi extension: clearer replies via Jev review + optional rewrite/visuals
questions: { is_urgent: boolean, department: choice, }
Opencode plugin using Jev (system one model) as part of software development process. Not affiliated with Opencode team.
questions: { alive: boolean, }
Jev, TypeSafe's System One classifier, as a tool inside Claude Code, Codex, Pi, and OpenCode: typed classify, check, score, rank, and ask, plus one-command setup.
questions: { outOfScope: boolean, contradictsPrevious: boolean, shouldFlag: boolean, }
Every tool call your agent makes, checked before it runs. A Claude Code plugin that uses TypeSafe AI's Jev to verify each pending tool call against the session plan, then allows it, asks you, or blocks it. Proof of concept
questions: { scope_alignment: score, touches_state: boolean, breaks_contracts: boolean, // +2 }
Fast semantic code search & diff sanity auditor for AI coding assistants (Antigravity, Cursor, Claude Code) powered by TypeSafe System One.
questions: { edit: choice, contract: choice, caller: choice, // +2 }
Semantic code review with Jev, plain-English rules and Agent Skills.
questions: { needs_vision: boolean, needs_tools: boolean, }
Pick the best AI model and reasoning effort for any task in ~1s. Plugin for Claude Code, Claude Desktop and Codex, powered by TypeSafe's Jev decision model and live OpenRouter pricing. Balance intelligence, speed and cost, or choose your priority.
questions: { disposition: choice, risk: score, data_exfil_risk: boolean, // +2 }
Agent tool/MCP call gate — allow / ask_human / deny via TypeSafe Jev
questions: { relevance: choice, }
Codex code-search plugin using Jev relevance filtering with auditable token metrics
from typesafe_sdk import AsyncTypeSafeClient, RetryPolicy
Software factory foreman based on TypeSafe's Jev model
import { choice, score } from "@typesafe-ai/sdk";Route to the cheapest model in claude code for your task using jev-router
? "https://api.typesafe.ai/v1/systemone"
Self-hosted, versioned skills library for AI agents. MCP, scoped clients, and optional Jev recommendations.
export const JEV_API_ENDPOINT = "https://api.typesafe.ai/v1/systemone";
Local-first MCP plugin for continuous software-quality review by AI coding agents, powered by Jev.
import { choice, noul, score } from "@typesafe-ai/sdk";Fast, cheap, typed judgments from TypeSafe's Jev model, as MCP tools.
"name": "jev",
A skill for writing and improving programs that call Jev, TypeSafe's System One model
if r.URL.Path != "/v1/systemone" || r.Header.Get("Authorization") != "Bearer k" {mcp connector to give your AI agent direct access to typesafe ai's jev model
export const SYSTEM_ONE_URL = 'https://api.typesafe.ai/v1/systemone';
Claude Code plugin: trim long Bash output with TypeSafe Jev before the model sees it
import { TypeSafeClient, choice, noul, type ChoiceCriteria, type EntryType } from "@typesafe-ai/sdk";🇺🇸 English · 🇧🇷 Leia em português
//! port, empty or root path only. Canonical origins append `/v1/systemone`
Rust CLI powered by Jev from TypeSafe.ai that ranks agent skills for the next step using live session context. Includes Claude Code hooks, structured JSON, abstention, and local feedback. Requires a TypeSafe API key.
"""One small, strict client for TypeSafe Jev (POST /v1/systemone).
Jev-powered model routing, memory, compaction, skill selection, computer and browser use for Hermes agents (also Claude Code and Codex)
baseUrl: process.env.JEV_BASE_URL || 'https://api.typesafe.ai/v1/systemone',
Portable, Jev-guided context compaction for coding agents.
import { createLiveTypeSafeClient, type TypeSafePort } from "../semantic/typesafe-client.js";CLI that picks Cursor, Claude Code, Codex, or OpenCode + model/effort for a task, then launches it. Powered by Jev and Herdr
from typesafe_sdk import AsyncTypeSafeClient, TypeSafeError
Typesafe.ai System One Model Jev navigating a Neo4j graph by using a classifier over neighbouring relationships
url: "https://api.typesafe.ai/v1/systemone",
Jev picks which of your rules apply to each prompt, so Claude only sees the ones that matter.
ENDPOINT = "https://api.typesafe.ai/v1/systemone"
Skill for Hermes, and other agents, to ask typesafe's jev
const validUrl=url.protocol==='http:' && url.hostname==='127.0.0.1' && /^\d+$/.test(url.port) && url.pathname==='/v1/systemone' && !url.username && !url.passwor
TypeSafe Jev action selection inside Codex Computer Use
TypeSafeClient,
Jev (TypeSafe System One) backed auto mode for the Pi coding agent: semantically auto-approves bash, write, and edit tool calls and fails closed when a decision cannot be made.
baseURL: "https://api.typesafe.ai",
原版 Pi Coding Agent 插件:按时机配置规则,并自带风险检查、输出脱敏、重复失败和缺少验证提醒。
endpoint: "https://api.typesafe.ai/v1/systemone".to_string(),
100% free ₹0 agent-first SEO & GEO CLI suite and MCP server in Rust replacing Semrush and OpenSEO via DuckDuckGo and TypeSafe Jev System One
logging.getLogger("typesafe_sdk").setLevel(logging.WARNING)MCP server for TypeSafe Jev: typed classify, score, check, match and screen for any agent, with confidence on every answer
id: "jev", name: "Jev (TypeSafe AI)",
Monitoring & Safety layer for all your agents. Open Source CLI & Skills for Claude Code, Codex, Cursor, Jev and your preferred agents.
endpoint = "https://api.typesafe.ai/v1/systemone"
A claude code plugin for jev
ENDPOINT = "https://api.typesafe.ai/v1/systemone"
An agent skill to discover TypeSafe Jev opportunities, design typed questions, and learn from recent community experiments.
"jev",
⚡ Sub-100ms cognitive reflexes for autonomous coding agents. Powered by TypeSafe AI's Jev & get-fable.
"base_url": "", # default https://api.typesafe.ai
TypeSafe (Jev) skill routing for Hermes Agent: names the one skill worth loading, before the model call. Opt-in, stdlib only, ~$0.001 per routed turn.
_ENDPOINT = "https://api.typesafe.ai/v1/systemone"
Typed, confidence-aware agent skill routing with TypeSafe Jev.
TYPESAFE_BASE_URL = "https://api.typesafe.ai"
Typed System One decisions, ranking, verification, and an opt-in Hermes tool gate using TypeSafe Jev.
'provider': provider, 'base_url': 'https://ai-gateway.vercel.sh/typesafe' if vercel else 'https://api.typesafe.ai',
Linux, macOS ve Windows için kaynaklı ikinci beyin. Claude Code, Codex ve Antigravity adaptörleri; yerel Markdown kasa, ayrı hafıza incelemesi, isteğe bağlı Mem0/Jev.
ENDPOINT = "https://api.typesafe.ai/v1/systemone"
用 TypeSafe Jev 推荐已安装 Skill / Bounded installed-skill recommendations with TypeSafe Jev. Python CLI, Codex skill, bilingual docs and live examples.
base_url="https://api.typesafe.ai", retry=RetryPolicy(max_retries=1),
TypeSafe AI Jev judgments for Agent Zero, with typed tools and probability cards.
} from "@typesafe-ai/sdk";
Hybrid coding harness: System 2 writes, System 1 (Jev) runs reflexes.
model: undefined, // undefined = advocaat default (jev-latest / typesafe-ai/jev)
Turn your AGENTS.md preferences into a fast, Jev-powered AI linter.
TYPESAFE_ENDPOINT = "https://api.typesafe.ai/v1/systemone"
Hand the browser work off: an MCP server where a decision model drives the page for your agent, so a flow costs one tool call instead of a turn per click. Ref-based element tables, code-checked assertions, zero-model macro replay. Speaks CDP to your Chrome.
"--provider", choices=["jev", "haiku", "anthropic"], required=True
Experimental Jev permission gate for Claude Code via OpenRouter, with reproducible latency and cost benchmarks
return (env.get("TYPESAFE_BASE_URL") or "https://api.typesafe.ai").rstrip("/")if you're experimenting with jev it will be easier from here
for u in "https://docs.typesafe.ai/llms.txt" "https://evals.typesafe.ai/" "https://openrouter.ai/typesafe/jev-1.13"; do
Agent skill: design judgment-assisted systems with TypeSafe Jev (System One). Maps Choice/Score/Noul onto decision theory, reranking, and routing. Composition algebra, question design, validation gates. MIT.
Thin httpx wrapper around POST https://api.typesafe.ai/v1/systemone.
TypeSafe Jev (System One) decision tools for Hermes Agent: jev_check / jev_route / jev_score / jev_evaluate
This is a thin pass-through for ``POST /v1/systemone``. The request body
Agent Skill: send closed coding-agent judgments to TypeSafe Jev
export const SYSTEM_ONE_URL = 'https://api.typesafe.ai/v1/systemone';
Codex plugin: verbatim Jev-guided context restoration around session compaction. Port of tamaratran/fast-jev-compaction to Codex lifecycle hooks.
export const JEV_URL = process.env.CANNY_JEV_URL ?? "https://api.typesafe.ai/v1/systemone";
Stops AI coding agents from claiming work is done without evidence. Deterministic hooks decide, TypeSafe's Jev advises. Append-only ledger, zero runtime dependencies.
return { kind: "connection", message, retryable: true, hint: "Check network access to api.typesafe.ai." };MCP server exposing TypeSafe Jev as typed, calibrated judgment tools: classify, score, check, batched ask. Ships as a Claude Code plugin.
process.env.TYPESAFE_ENDPOINT ?? "https://api.typesafe.ai/v1/systemone";
Reward-hack radar for coding agents: structural denies + TypeSafe Jev System One sidecar for Claude Code & Cursor hooks
from typesafe_sdk import Choice, Noul, TypeSafeClient
Natural-language MCP tool dispatcher powered entirely by TypeSafe's Jev — no general-purpose LLM. Discovers a simple MCP server's tool signatures at runtime and uses Jev's typed primitives (Choice/Noul) to pick the right tool and extract its arguments straight out of the sentence.
import { noul, TypeSafeClient, type NoulResponse } from "@typesafe-ai/sdk";Plugin for oh-my-pi that uses Typesafe Jev API to classify tool calls as safe/unsafe/ask
_TOOLSET = "jev"
Jev Decisions Plugin for Hermes (and other AI Agents): tool risk reviews, human approval recommendations, evidence checks, and a local decision journal.
import type { EntryType, NoulQuestion as SdkNoulQuestion, Usage } from '@typesafe-ai/sdk';Claude Code plugin that scores how well you prompt a coding agent, and shows whether your habits are improving. Runs on TypeSafe's Jev model. Zero added latency.
DEFAULT_ENDPOINT = "https://api.typesafe.ai/v1/systemone"
I kept watching coding agents burn context on decisions that aren't hard - triage 400 tickets, tag 600 files, route to one of six teams. jev-mode moves those verdicts to a typed-judgment model. I A/B'd it: 78% fewer tokens, 16x less work-attributable input, accuracy 96.1% vs 93.7%. Python, no deps, MIT.
// POST /v1/systemone {model, state, questions} -> {model, answers, usage}TypeSafe AI System One (Jev) task plugin for QuantumNous/new-api — native /v1/systemone, synchronous evaluation, token billing
request = Request('https://api.typesafe.ai/v1/systemone',Copy and paste this into your coding agent:
TypeSafeClient,
Select Git changes for staging with a plain-language description.
/** Defaults to `~typesafe/jev-latest`. */
Fork of tamaratran/fast-jev-compaction: Jev via OpenRouter with zero data retention (zdr, data_collection: deny)
pub const DEFAULT_API_URL: &str = "https://api.typesafe.ai";
Rust-aware code review for Claude Code and coding agents, powered by TypeSafe Jev
from typesafe_sdk import Choice, Noul, Score, TypeSafeClient, TypeSafeAPIError
Discriminative Monte Carlo Tree Search using TypeSafe Jev System One Primitives and Gemini
from typesafe_sdk import Choice, Noul
Tell Claude Code which installed skill a session needs, using Jev (TypeSafe AI) for the decision and skills.sh for discovery.
TYPESAFE_URL = "https://api.typesafe.ai/v1/systemone"
MCP server: screen PubMed titles/abstracts against a query or clinical question with TypeSafe Jev
const ORIGIN = "https://api.typesafe.ai";
Codex 可恢复委派:Capsule → Jev 选路 → Policy Guard(仅 ALLOW/DENY)→ Root 验收。自动委派默认关闭。Recoverable Codex delegation via Jev + Policy Guard; Root keeps acceptance.
export const JEV_ENDPOINT = 'https://api.typesafe.ai/v1/systemone';
Not every coding task needs your best model. Experimental Jev-powered model routing for Claude Code — V3 prototype runs today, V4 routes at the task boundary.
with patch('jevseek.models.OpenAI') as ds, patch('jevseek.models.TypeSafeClient') as jev:A local coding workspace pairing Jev action routing with DeepSeek argument generation. Native tools, persistent sessions, React desktop, and documented research.
import { experimental_evaluate as evaluate } from 'ai';Jev-backed tool approval gate and tool-list pruning for the Vercel AI SDK
export const JEV_MODEL = '~typesafe/jev-latest'
星露谷农场小助手:Jev 自主游玩、dsh 插件、独立 CLI 与 SMAPI Mod
baseUrl: "https://api.typesafe.ai/v1",
Cut Claude Code's skill manifest by ~75% with TypeSafe Jev. Scores every installed skill for relevance and hides the rest via skillOverrides — 12,750 → 3,185 tokens on a 217-skill install, for $0.0009 a session.
API = "https://api.typesafe.ai/v1/systemone"
Claude Code Stop hook that checks an AI assistant's claims against what it actually read this session, using TypeSafe's Jev as the judge
const API_ENDPOINT = process.env.TYPESAFE_ENDPOINT || "https://api.typesafe.ai/v1/systemone";
Universal Model Context Protocol (MCP) Server for TypeSafe Jev (System One) semantic code search and validation.
server = mcp.get("mcpServers", {}).get("jev", {})Codex Plugin with typed Jev judgments for risk review, evidence checks, context screening, and reranking.
import { choice, score } from "@typesafe-ai/sdk";Keeps OpenCode on a cheap sticky model for warm cache; Jev escalates hard turns to stronger subagents.
let response = match ureq::post("https://api.typesafe.ai/v1/systemone")Zero-hallucination open-source repo and crate scout powered by TypeSafe AI Jev System One scoring
export const DEFAULT_BASE_URL = "https://api.typesafe.ai/v1/systemone";
Per-prompt capability router for coding agents: resolves installed skills, MCP servers, agents and commands against your prompt via TypeSafe Jev, and measures whether the injection actually helps.
API = "https://api.typesafe.ai/v1/systemone"
Скилл для агентов Letta: суждения по критериям через TypeSafe System One (Jev)
join(or(env.TYPESAFE_BASE_URL, 'https://api.typesafe.ai'), '/v1/systemone'),
Typed decisions in Claude Code: adds $.jev over TypeSafe's Jev, through OpenRouter, Vercel AI Gateway, Cloudflare Workers AI, LiteLLM or the TypeSafe API.
DEFAULT_MODEL = os.environ.get("JEV_MODEL", "typesafe/jev-1.13")Jev for Hermes: cheap intent gates + verbatim tool compaction on OpenRouter
* SDK's `experimental_evaluate`, not the chat-completions endpoints.
What your last session knew, scored against what this one is doing. MCP server: a per-project ledger written as things happen, recalled per task with TypeSafe's Jev evaluation model via Vercel AI Gateway.
"""Native HTTP client for Experiential /v1/systemone.
A gate for your agent's expensive steps, powered by TypeSafe Jev (System One). Offline-first, OpenRouter or direct, MIT.
typesafe: { key: "TYPESAFE_API_KEY", model: "jev-latest", endpoint: "https://api.typesafe.ai/v1/systemone" },Local tool-routing classifier for coding agents, with a gateway, MCP integrations, and decision logs.
const defaultModel = "typesafe/jev-1.13";
Stop burning LLM calls on classification. Route bugs, triage failures, and gate PRs in 200ms for $0.00002. OpenClaw plugin for TypeSafe Jev structured decisions.
ENDPOINT = os.environ.get("TYPESAFE_API_BASE", "https://api.typesafe.ai") + "/v1/systemone"Agent skill that spots bounded-judgment steps, tries a typed decision model (TypeSafe's Jev) first, and documents every attempt
const typesafe = configFromEnv({ TYPESAFE_API_KEY: 'ts', JEVC_MODEL: 'jev-1.13', JEVC_BASE_URL: 'https://proxy.example/v1/systemone' });Pi extension suite powered by Jev: selective context compaction and model routing
"""Thin HTTP client for the TypeSafe /v1/systemone endpoint.
Everything you need to run TypeSafe's Jev with Claude Code: a tool-call guard, tier guard, file search, browser agent, review, belay, compaction and installers.
TYPESAFE_BASE_URL: Final = "https://api.typesafe.ai"
Home Assistant custom component: Conversation agent with Jev fast-path + Grok fallback
const API_URL = 'https://api.typesafe.ai/v1/systemone';
Auto permission mode for Command Code: screens every tool call with TypeSafe Jev before it runs, and rejects anything out of scope.
API_URL = "https://api.typesafe.ai/v1/systemone"
Jev powered code review hook for agents
const Endpoint = "https://api.typesafe.ai/v1/systemone"
A bounded Jev risk check for Claude Code: eight risk axes, one request, one optional reinspection.
const response = await (options.fetch ?? fetch)("https://api.typesafe.ai/v1/systemone", {pi-jev-agent is an experimental extension for the Pi coding agent. It uses TypeSafe Jev to choose the next tool before each language-model step.
// "~typesafe/jev-latest" always points to the newest Jev. The tilde matters:
Throw in a pile of company files and get them classified and organized by department, type, sensitivity, date, counterparty and PII, with an index for AI agents. Powered by TypeSafe's Jev on OpenRouter (17¢ per 1,000 files). Zero-dependency Node CLI + Claude skill + Codex agent.
"typesafe": ("https://api.typesafe.ai/v1/systemone", "jev-latest", "noul", "noul"),A Stop hook that stops your coding agent from stopping too early. Plain-language rules, judged by jev.
TS_URL = os.environ.get("TS_MCP_TS_URL", "https://api.typesafe.ai/v1/systemone")MCP server exposing TypeSafe (Jev/System One) to the fleet: judge, rerank, systemone
# - TypeSafe native-shaped: POST /v1/systemone, GET /v1/models
Typed decisions from the shell: an unofficial stdlib-Python CLI and Agent Skill for TypeSafe's Jev model, via the TypeSafe API (default) or OpenRouter. Yes/no, choice and ordinal scores with calibrated probabilities, semantic grep and batch mode.
const fn = (mod['experimental_evaluate'] ?? mod['evaluate']) as EvaluateFn | undefined;
Busca em arvore (MCTS/PUCT) com avaliacao tipada do TypeSafe Jev e portao humano obrigatorio. A arvore supoe, a sonda mede: marco so e concedido por codigo de saida verde.
DEFAULT_URL = "https://api.typesafe.ai/v1/systemone"
Hermes Agent skill whose north-star gate is judged by Jev (TypeSafe System One): turn an intention into a checkable finish line, generate the run prompt, and let Jev rank what is still unproven.
export function createDecider(endpoint = "https://api.typesafe.ai/v1/systemone"): Decide {Opt-in Auto (Jev) model routing for OpenCode with a configurable model allowlist
TYPESAFE_URL = "https://api.typesafe.ai/v1/systemone"
general Chrome agent — Jev + OpenRouter
model: "~typesafe/jev-latest",
open-sourced jev-flash-router: an MCP server for TypeSafe's new Jev model. AI coding agents waste hundreds of reasoning tokens just deciding which file to edit, which route to pick, or whether a diff breaks tests. Jev evaluates state and outputs calibrated probabilities. Works with Cursor, Windsurf, & Claude Code
const DEFAULT_API_BASE = "https://api.typesafe.ai/v1";
Pi extension that classifies each prompt with TypeSafe's Jev and routes it to the best model you're logged into.
model: str = "typesafe:jev-1.13.0"
TypeSafe/Jev router plugin for Hermes Agent — compact tool results, suppress duplicate tools, skip unnecessary main-model calls
import { type NoulQuestion, noul, TypeSafeClient } from "@typesafe-ai/sdk";Fast deep research for the pi coding agent: reads up to 100 pages in full per round, Jev keeps only the passages that answer your questions.
import { TypeSafeClient, type TypeSafeClientConfig } from '@typesafe-ai/sdk';Analyze prompts with TypeSafe Jev before GitHub Copilot
TypeSafeClient,
TypeSafe AI (Jev) adversarial reviewer and typesafe_ask tool for the omp coding agent
TypeSafeClient,
Computer Use Agent developed with Jev
const response = await (this.options.fetch ?? fetch)('https://api.typesafe.ai/v1/systemone', {Jev-powered tool and skill selection, context search, and output triage for Codex via MCP
response = await request('https://api.typesafe.ai/v1/systemone', { method: 'POST',A small advisory tool for comparing supplied alternatives against supplied evidence.
from langchain_typesafe import Choice, Noul, NoulCriteria
LangChain / Deep Agents middleware that uses TypeSafe's System One model (Jev) for typed judgments in unattended coding agents: a shell-command gate (database, production, destructive, secrets), issue triage and routing by severity and urgency, merge-request detection, and review of weakened tests. Measured with live probes.
DEFAULT_MODEL = "~typesafe/jev-latest"
A Hermes Agent model-provider plugin that routes selected auxiliary tasks through bounded Jev decision calls on OpenRouter instead of free-form chat prompts.
constructor({ apiKey = process.env.TYPESAFE_API_KEY, endpoint = process.env.TYPESAFE_ENDPOINT ?? "https://api.typesafe.ai/v1/systemone", model = process.env.TYPPortable System-1 decision layer for agent harnesses with host-owned routing, receipts, replay, and fail-open integrations.
"jev",
Portable agent skill: TypeSafe Jev as a cheap code-review classifier (HTTP + optional jev-review MCP)
questions to typesafe_sdk types and projects the response back to jcyber's
Agent-driven bug bounty / pentest framework: one gated chain over five systems (Caido, HexStrike, Jev, Memgraph, TencentDB, Prometheus)
name="jev",
Give any MCP-capable LLM harness an on-demand real-browser search tool (TypeSafe Jev) with per-run timing and cost tracking.
"model": os.environ.get("JEV_DECISIONS_MODEL", "~typesafe/jev-latest"),jev-mcp turns browser-use/jev-ultrafast into one local, session-aware MCP server for fast read-only product research. Jev selects an action and observed target from a DOM snapshot;
export const ENDPOINT = 'https://api.typesafe.ai/v1/systemone';
Jev-powered context curation for Codex. Build compact, traceable handoff context through native plugins and skills.
"3. Would a sensible person answer it in under a second from text you can show them? -> TypeSafe/Jev.\n" +
A Claude Code hook that asks whether the decision you are writing needs a model at all. Includes a measured 149-row comparison of TypeSafe Jev against Claude Haiku 4.5.