effortless
Claude Code mod that picks reasoning effort per prompt with a Haiku judge, optionally Jev, and shows cache and context state.
Overview
effortless is a Claude Code mod that chooses the reasoning effort for every prompt, so quick questions get Low and hard jobs get High without touching the effort control. The README reports that on real coding tasks it cut Opus cost by 21% against the recommended medium effort, 29% against high and 42% against xhigh, with every answer still correct. A bar above the prompt also shows how full the chat is and how long the prompt cache stays warm, with one-click hand-off or compaction.
By default the judge is Haiku 5.5 on your own Claude login, taking about a second. With a TypeSafe key, Jev answers the effort first in about 0.25 seconds and Haiku steps in when Jev is unsure. Haiku also summarises compactions, and an experimental, off-by-default option can run simple prompts on a cheaper model, which the README explains often costs more because of separate prompt caches.
What the decision model decides
| Type | Question | Answers |
|---|---|---|
| Choice | Which reasoning effort does this prompt need?(inferred) | Low · Medium · High |
How it works
- User submits a prompt in Claude Code
- Jev (if configured) picks the effort; Haiku handles uncertain cases or runs alone
- The prompt runs at the chosen effort
- Status bar shows context fill and cache warmth; hand off or compact on demand
Strengths
- Reported Opus cost reductions with correct answers
- Works without a TypeSafe key via Haiku
- Jev makes judging about 4x faster
Caveats
- Jev is optional; default judge is a generative model
- Benchmark is the author's own
Getting started
In Claude Code: claude plugin marketplace add HeyCubit/effortless, then claude plugin install effortless@effortless and /reload-plugins.At a glance
- Type
- Project
- Status
- beta
- Language
- TypeScript
- Runtime
- Claude Code plugin/mod
- Has code
- yes
Use cases and integrations
How curated lists describe it
Claude Code mod that picks the reasoning effort for every prompt, shows prompt-cache and context state, and hands off or compacts in one click; Haiku by default, Jev with a TypeSafe key.
MrJev/awesome-jev
Frequently asked questions
Do I need Jev?
No, Haiku judges by default; adding a TypeSafe key makes Jev the faster first judge.
How much does it save?
The README reports 21% to 42% lower Opus cost depending on the baseline effort.
Does it switch models?
Only with an experimental, off-by-default setting.
Related projects
Claude Code plugin that picks model tier and effort per turn with Jev and weighs each switch against cost, with Jev-based compaction.
ProjectbetaTypeScriptClaude Code and Codex plugin where Jev judges each turn for model routing, delegation, evidence search and compaction.
ProjectbetaTypeScriptClaude Code mod using Jev to pick model and effort per turn, size subagents, and prune stale context verbatim instead of summarising.
ProjectbetaTypeScriptOpen-source AI gateway for 100+ LLMs that can use Jev to classify requests for complexity routing and to compact tool results.
ProxyproductionPythonHermes Agent plugin that picks the model each turn with a direct TypeSafe Jev Choice and enforces trim-then-compress context handling.
PluginbetaPython- 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