# Jev with AI agents

> Setup guides for Jev with Hermes, OpenClaw, Claude, Codex, Cursor, Copilot, LangChain, Cline, Roo, Devin, and browser agents.

Let the chat model write prose. Let Jev decide which shell command, skill, route, or gate fires next.

Generative agents are brilliant at language and terrible at pretending to be a calibrated classifier. Jev answers fixed-shape questions with probabilities you can threshold in code. These guides map verified integrations from the ecosystem: Python harness middleware, reflex MCP servers, plugins, and agent skills that call POST /v1/systemone (or gateway equivalents) while your main model keeps the microphone.

## Guides

- [Jev with Hermes Agent](https://jev.aitools.fyi/guides/jev-with-ai-agents/hermes): Hermes Jev means Jev reviews flagged Hermes shell commands as APPROVE, DENY, or ESCALATE, so Hermes can quit faking a classifier.
- [OpenClaw + Jev](https://jev.aitools.fyi/guides/jev-with-ai-agents/openclaw): Layer OpenClaw plugins jev-claw and jev-harness, plus bundled typesafe/jev-latest decisionModel. Tools, config, and when to route vs gate.
- [Claude Code, Desktop, and Cowork + Jev](https://jev.aitools.fyi/guides/jev-with-ai-agents/claude-code-desktop-cowork): Wire Jev into Claude Code and Cowork with typesafe-ai/skills, jev-judge-mcp, typesafe-mcp, and jevwire. Keys, install commands, and Cowork notes.
- [Codex and OpenCode + Jev](https://jev.aitools.fyi/guides/jev-with-ai-agents/codex-and-opencode): Use Jev with Codex and OpenCode to judge tool calls and pick a route. MCP setup, skills, and a router that fails open if Jev errors.
- [Cursor, Grok, and agent skills + Jev](https://jev.aitools.fyi/guides/jev-with-ai-agents/cursor-mcp-and-skills): Add jev-mcp to Cursor MCP config, install typesafe-ai/skills for Cloud Agents, and use Jev beside Grok or other models without custom Grok Bot APIs.
- [LangChain and LangGraph + Jev](https://jev.aitools.fyi/guides/jev-with-ai-agents/langchain-langgraph): Install langchain-typesafe and use Choice, Score, and Noul with TypeSafeClassifier, plus router or auto-mode middleware. Patterns, pitfalls, and catalog links.
- [Cline and Roo Code + Jev](https://jev.aitools.fyi/guides/jev-with-ai-agents/cline-and-roo): Wire Jev into Cline and Roo via system1-mcp uvx install, itsmostafa/typesafe-mcp, or jev-mcp. cline_mcp_settings.json, tool surfaces, env keys, and pitfalls.
- [GitHub Copilot Agent Mode + Jev](https://jev.aitools.fyi/guides/jev-with-ai-agents/github-copilot-agent): Jev in VS Code is an MCP server you start and trust, then use in Copilot Agent mode. .vscode/mcp.json uses servers. Copilot CLI uses mcpServers.
- [Devin Desktop + Jev](https://jev.aitools.fyi/guides/jev-with-ai-agents/devin-desktop): Pair Jev with Devin Desktop (ex-Windsurf) via system1-mcp MCP install, multi-agent ACP surfaces, and catalog MCP listings. No invented native plugins.
- [Browser and computer-use agents + Jev](https://jev.aitools.fyi/guides/jev-with-ai-agents/browser-computer-use): Pair Jev with Browser Use and computer-use stacks: next-action Choice, risk nouls, and demos for ultrafast browser use and typesafe-computer-use.

## FAQ

### Do I need a separate API key for every agent?

Usually one TypeSafe key is enough. Export TYPESAFE_API_KEY in the shell or service that launches the agent (or reference it from MCP config with env interpolation where supported). Vercel AI Gateway users can use AI_GATEWAY_API_KEY for gateway-routed Jev instead; see the learn guide on Vercel AI Gateway Jev.

### MCP server or agent skill: which comes first?

Skills teach the model when and how to design a Jev call. MCP servers expose tools like judge, evaluate, or jev_route that actually hit the API. Many teams install both: typesafe-ai/skills for patterns, plus a focused MCP package for the hot path.

### What happened to system1-mcp from Nous Research?

We did not find a public NousResearch/system1-mcp repository. The PyPI package system1-mcp (ericmaddox/system1-mcp) is a separate community reflex MCP server with fast_guard, fast_judge, fast_verify, and fast_score. Hermes users can still use Hermes plugins, generic MCP, blakestone-x/jev-mcp, or itsmostafa/typesafe-mcp.

### LangChain middleware vs MCP reflex tools: which first?

Use langchain-typesafe when you own the Python agent graph and want TypeSafeClassifier, ModelRouterMiddleware, or AutoModeMiddleware inside LangChain. Use system1-mcp or typesafe-mcp when the host is an IDE agent (Cline, Copilot, Cursor) that only speaks MCP. Many teams do both: middleware in production services, MCP in the editor.

### Is there a Muse plugin that uses Jev?

We did not find a documented OpenClaw or TypeSafe integration branded Muse with Jev-specific tools. If you discover one, submit it to the directory. Until then, treat Muse as unrelated to Jev unless the upstream README names TypeSafe or System One explicitly.

### Where do Choice, Score, and Noul show up in agents?

Approvals and routing use Choice (pick APPROVE/DENY/ESCALATE or a model route). Risk and severity use Score. Safety checks and policy fit use Noul (0 to 1). Parallel questions in one request are the default System One pattern; see /learn/system-one.

### Can Jev replace my main chat model?

No. Jev does not draft user-facing paragraphs or long reasoning chains. It gates, ranks, classifies, and scores. Keep Claude, GPT, Grok, or Hermes' primary model for language; add Jev where you would otherwise regex-parse YES/NO from a completion.

Directory copy checked 2026-09-30.
Canonical: https://jev.aitools.fyi/guides/jev-with-ai-agents