jev-model-router (Mandrilsquad1441)
Claude Code, Claude Desktop and Codex plugin where Jev reads a task and ranks live models and reasoning effort by quality, speed and cost.
Overview
Jev Model Router is a plugin and MCP server for Claude Code, the Claude desktop app and Codex that recommends which model and reasoning effort to use for a task. Jev reads the task in about 0.4 s and classifies what kind of work it is, how hard, how much reasoning it needs, and whether it needs tools, vision, real-time answers or high volume.
Live models are then ranked in code against those needs using benchmark-based quality, live OpenRouter prices, speed and suitable effort. The output is a best-balance pick plus smartest, fastest and cheapest good-enough alternatives, with the exact command to switch, such as /model and /effort in Claude Code or codex flags, or an OpenRouter request body. The README reports a cost analysis on 42 real labelled developer tasks at live list prices, reproducible with a script.
What the decision model decides
| Type | Question | Answers |
|---|---|---|
| Choice | What kind of work is this task (for example debugging)? | — |
| Score | How hard is this task (0-4)? | — |
| Yes / no | Does it need tools, vision, real-time answers or high volume? | yes · no |
How it works
- The user asks which model to use for a task
- Jev classifies task type, difficulty, reasoning need and signals
- Code ranks live models by quality, price, speed and effort
- The plugin returns a pick, alternatives and the switch command
Strengths
- Uses live OpenRouter pricing
- Gives exact commands to switch model and effort
- Cost analysis on labelled tasks is reproducible
- CI and MIT licence
Caveats
- Quality scores come from benchmarks, not your workload
- Requires a TypeSafe key
- Savings figures depend on list prices at the time
At a glance
- Type
- Plugin
- Status
- beta
- Language
- TypeScript
- Runtime
- Claude Code / Codex plugin and MCP server
- Has code
- yes
- Has tests
- yes
- License
- MIT
Use cases and integrations
How curated lists describe it
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.
hellogumbo/awesome-jev
Frequently asked questions
What does Jev decide?
The task's type, difficulty, reasoning need and signals such as tool use or vision.
Does it switch models automatically?
No. It recommends a pick and gives the exact command to switch.
Which clients does it support?
Claude Code, the Claude desktop app and Codex.
Related projects
- jev-router★ 593
Per-turn model routing for Claude Code and Codex: wraps the real CLI and lets Jev send simple turns to fast models and hard ones to strong.
ProxybetaJavaScript TypeScript/Bun launcher that asks Jev which model and reasoning effort each Codex, Claude Code or Pi turn should use.
ProjectbetaTypeScriptGo CLI model router that uses one Jev analysis to recommend a model, harness and effort level, then selects deterministically.
CLIbetaGoClaude Code and Codex plugin where Jev judges each turn for model routing, delegation, evidence search and compaction.
ProjectbetaTypeScriptLocal pass-through proxy for Claude Code and Codex where Jev picks a model tier per human turn and only moves up within a session.
ProxybetaTypeScriptLocal OpenAI/Anthropic-compatible proxy where one Jev call picks the model route and thinking level for each coding-agent turn.
ProxybetaGo