robertn702/opencode-jev-routerArchived
User repository on GitHub
Superseded by robertn702/reasoning-router
- Forks
- 0
- Last pushed
- Updated Oct 2026
Repo stats from the GitHub API, cached Oct 2026.Homepage
Jev directory
CommunityJev Integrations & Adapters
Opencode Jev Router
An OpenCode plugin that asks Jev how much reasoning each step needs — low for a rename, high for a failing test — and adds the picked effort to the request, falling back to high when Jev is slow or unavailable. Archived and superseded by reasoning-router.
- Stars
- 10
- Language
- TypeScript
- License
- MIT
- Reading time
- 2 min
- Published by
- Community
- Author
- robertn702
- Last verified
- Oct 2026
Works forJev for Routing
Used inJev for Software & DevTools
- Languages
- TypeScript
- Platforms
- CLIVercel AI Gateway
- Capabilities
- Open sourceRequires API key
Highlights
- Archived: superseded by robertn702/reasoning-router; the plugin is now @reasoning-router/opencode.
- Adds the chosen effort as an OpenAI configuration_update item; earlier updates stay in place so the prompt prefix stays cacheable.
- Reported over the tested GPT-6 Astra mix: both Jev and fixed high solved 44/44 attempts, Jev using 14% fewer output tokens.
- On pytest #5262 it used 42% fewer output tokens and 40% less time, both solving 6/6; a separate hard case went 5/5 versus 1/5.
- Needs OpenCode V2 2.0.4+, a Jev key from TypeSafe or Vercel AI Gateway, and a Responses API-compatible endpoint.
Quickstart
bash
export JEV_API_KEY=...
export OPENAI_API_KEY=...Watch out
Grounded on a truncated README and GitHub metadata captured 2026-10-15; the repository is archived and its own banner says it is superseded by reasoning-router, so new setups should use the successor. The evaluation numbers are the project's own and the README notes results on other workloads may differ.
Questions about this entry
Short answers drawn from the linked project or page.
It keeps the quality and spends less reasoning: for each primary request to a wrapped model it asks Jev how much reasoning the step needs, then adds that effort to the request so one model handles quick edits and hard debugging.
No. The repository is archived and its README says it is superseded by robertn702/reasoning-router; the OpenCode plugin is now @reasoning-router/opencode. Use this entry for reference and the successor for new setups.
The plugin sends a bounded summary of the recent conversation to Jev, which picks a reasoning effort. That choice is added to the request as an OpenAI configuration_update item, or as an Anthropic effort-only system message, and the response streams back unchanged.
Across the tested GPT-6 Astra task mix, Jev and fixed high effort each solved 44/44 attempts; Jev used 14% fewer output tokens including reasoning and finished 9% faster on average. The README notes results on other workloads may differ.
OpenCode V2 2.0.4 or newer, a Jev key from TypeSafe or Vercel AI Gateway, and a Responses API-compatible endpoint that serves GPT-6 Astra, Luna or Sol and accepts configuration_update items, plus its API key.
If Jev is slow or unavailable, the request continues at a fallback effort, high by default. Each routed request makes one classification call, billed by the Jev provider.
The README gives low for a rename and high for a failing test as examples. Earlier updates stay in place so the prompt prefix stays cacheable, and the chosen effort is applied mid-conversation.
More Jev resources like this
An opencode plugin that turns an agent's small decisions — is this command safe, which tool fits, is this PR ready — into typed Jev questions with calibrated confidences, labelling each answer auto or escalate so the agent knows when to ask for help.
Jev Integrations & Adapters#community#catalog-import#logicrw
Community
darwintechlab/openjevJavaScript
darwintechlab/openjevJavaScriptAn OpenCode plugin that keeps the session on a cheap sticky parent model so the cache stays warm, and asks Jev whether a turn is hard before escalating through jev_escalate to temporary stronger subagents.
Jev Integrations & Adapters#community#catalog-import#cobanov
A Claude Code plugin that picks the model tier and thinking effort for every turn with Jev and weighs each switch against what it costs, since moving a long conversation to another model rewrites its prompt cache and can cost more than it saves.
Jev Integrations & Adapters#community#catalog-import#logicrw
A host-agnostic routing layer that sits between a coding harness and its models, choosing capacity per round from trajectory evidence like tool calls, failures and context pressure, running as a user-scoped daemon and local OpenAI-compatible proxy.
Jev Integrations & Adapters#community#catalog-import#logicrw
Community
vizuh/sabiTypeScript
vizuh/sabiTypeScriptLocal tool-routing classifier for coding agents: a gateway and MCP tool that ask Jev which tool to call next, keep your agent's own model and login, and log every decision so you can start by observing.
Jev Integrations & Adapters#community#catalog-import#logicrw
Community
felpsdev/jev-classifierTypeScript
felpsdev/jev-classifierTypeScriptAn OpenCode 2 plugin that sends every shell command — local or over FarHand's remote shell — to Jev first: six typed questions about global installs, host litter, harmful actions and privacy, and a permission prompt whenever something looks risky.
Jev Integrations & Adapters#community#catalog-import#logicrw
Community
cogflux/opencode-jev-guardTypeScript
cogflux/opencode-jev-guardTypeScriptFrom the Jev guides
Original write-ups on what this entry is for.
From the Jev AI model community
Posts from builders shipping with Jev right now.