Classification & Routing

Hermes Agent tool plugin giving bounded Jev judgments and an explicit model recommendation via TypeSafe, Cloudflare or OpenRouter.

Experimental Hermes plugin returning routing plans: Jev rates task complexity, deterministic policy picks the cheapest eligible model.

Experimental Hono router that matches HTTP requests to plain-language route descriptions using one Jev Noul per route.

Natural-language SQLite interface where code proposes options, Jev chooses among them, and a compiler builds deterministic SQL.

Local proxy for Claude Code that uses Jev to pick model and effort per message, routing subagents but guarding the cached main chat.

Colab notebook experiment that analyzes a movie review with one Noul, one Choice and one Score question to Jev.

Prototype per-call model router for Codex where Jev picks the GPT model and reasoning effort for each call via a local Responses proxy.

CLI, MCP server and skill for issue triage where Jev predicts issue kind, affected module and whether information is sufficient.

Local Anthropic-compatible proxy where Jev scores each Claude Code turn and simple work is routed to cheap or free models.

Two small Python apps: Jev triages failed pipeline tasks into review, retry, page or ticket, and routes support messages to agents.

Demo that classifies support tickets in BigQuery via a remote function backed by a Cloud Run adapter calling Jev.

French library ranking ADEME emission factors for an activity label, splitting code rules from Jev's semantic match.

Single-file Python proxy that asks Jev which model tier each Codex turn needs, rewrites the model, and escalates on failures.

Jev Router for Cursor: Jev picks a COST, BALANCED or INTELLIGENCE tier per task, with soft routing in chat and hard routing via the SDK.

Event-stream layer labelling signups, logins and checkouts as human, AI agent, crawler or bot, with a bad-intent probability.

Three Italian-language Claude Code prompts that build Jev-powered model routing, context compaction and code review into your project.

Web app that reads a two-person chat and labels each message's emotion, intent and reply quality to estimate interest.

Gateway that batches agent control questions through Jev and calls a generative LLM only when policy says generation is needed.

Undocumented app that uses Jev to classify text and pick an emoji reaction.

Experimental plugin wiring Jev into coding-agent harnesses to pick skills, tools, models and browser moves, starting with OpenCode.

Eight zero-dependency examples applying Jev to mechanical and electrical engineering: CAD routing, FEM triage, DFM, BOM alignment.

Self-hosted LLM gateway with semantic cache, privacy sandbox and a 'Jev classifier' tiering requests across cheap and frontier models.

Turns an inbox into a short action queue: seven Jev questions per thread decide what needs you, urgency and next step.

Python example classifying customer support messages into 12 intents with Jev, flagging clarification below 0.8 confidence.

Smart-meter fleet dashboard where code pre-filters telemetry and Jev classifies anomalies for remediation across millions of endpoints.

Read-only GitHub dashboard where Jev classifies candidate issue pairs as duplicate, related, distinct or underdocumented, with evidence.

Read-only Chrome extension that badges each Gmail message with Jev's category, priority, spam % and reply %.

Email triage proof of concept: one Jev call with 11 narrow questions per message, combined in Python into kind and topic labels.

Web chat where one Jev Choice picks nano, fast, balanced or frontier model tier per message via the Vercel AI Gateway.

Python router where Jev classifies prompt difficulty and stakes, then confidence-gated rules choose an OpenAI model tier.