openclaw-jev-plugin
OpenClaw plugin where Jev decides whether a group-chat bot should reply and, optionally, which model tier should answer.
Overview
openclaw-jev-gate is an OpenClaw plugin that makes two decisions before the language model runs. The reply gate, on by default, sits between requireMention true and false: the bot sees every group message but answers only those meant for it, such as mentions and follow-ups to its own replies, staying silent on chatter like lunch plans or thanks to a colleague. A silenced message costs one Jev call and no model tokens; direct messages are always answered unless you gate them too.
The optional model router asks Jev how much work a request needs and whether a wrong answer has consequences, then picks light, standard or heavy from models you configure. High-stakes requests go heavy, and when Jev is unsure the request goes to standard, never light. With no tiers set it only logs what it would do. The router works only for agents OpenClaw runs against a provider API, not CLI runtimes.
What the decision model decides
| Type | Question | Answers |
|---|---|---|
| Yes / no | Is this group message meant for the bot? | yes · no |
| Choice | Which model tier should answer? | light · standard · heavy |
How it works
- Message arrives in a group chat.
- Mentions always pass; otherwise Jev decides whether to reply.
- If routing is on, Jev rates effort and stakes.
- Select the tier (standard when unsure) and run the model.
Strengths
- Saves model tokens on messages not meant for the bot.
- Conservative routing never sends uncertain requests to light.
- Log-only mode for safe evaluation.
Caveats
- Router does not work with CLI runtimes.
- Requires a TypeSafe key.
At a glance
- Type
- Plugin
- Status
- beta
- Language
- TypeScript
- Runtime
- OpenClaw plugin
- Has code
- yes
Use cases and integrations
How curated lists describe it
A silenced message never reaches the language model, so it costs one Jev call and no model tokens. Direct messages always get an answer unless you choose to gate them too.
logicrw/awesome-jev-projects
Frequently asked questions
What does the reply gate do?
Lets the bot see all group messages but answer only the ones meant for it.
Where do uncertain requests go?
To the standard tier, never light.
Does it work with Claude CLI agents?
The router does not; it only works for API-backed agents.
Related projects
Transparent proxy that asks Jev to classify each coding-agent prompt and injects the decision into the agent's context before it generates.
ProxybetaPythonCost-, cache- and quota-aware LLM router that classifies each turn with a Jev-class model, prices every route, and checks cheap answers.
ProjectprototypePythonLocal-first routing engine for long-running agent tasks that uses a Jev decider to pick a model group per turn.
PluginproductionRust, TypeScriptOpen-source AI gateway for 100+ LLMs that can use Jev to classify requests for complexity routing and to compact tool results.
ProxyproductionPythonOpen-source workspace for running and reviewing AI coding agents across devices, with an optional Jev-based automatic model router.
ProjectproductionTypeScript- 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