User repository on GitHub
插件化 AI Agent 底座（FastAPI + Vue3）：多 Provider 模型接入与状态管理、Jev 结构化决策、插件市场一键安装、全局养成元气宠物、会话与插件一站式管理，内置一键启停脚本。
- Forks
- 3
- Last pushed
- Updated Oct 2026
Repo stats from the GitHub API, cached Oct 2026.
Jev directory
CommunityJev Agents & MCP
Andy Harness
A plugin-based agent harness (FastAPI + Vue 3) bundling a chat GUI, multi-provider model access, an MCP client and marketplace, skills, workflow studio, sandboxed execution and a Tauri desktop shell. It integrates TypeSafe AI's Jev for structured Choice, Score and Noul decisions.
- Stars
- 5
- Language
- TypeScript
- License
- MIT
- Reading time
- 3 min
- Published by
- Community
- Author
- andyqiuqiubo
- Last verified
- Oct 2026
Used inJev for Software & DevTools
- Languages
- PythonTypeScript
- Capabilities
- Open sourceSelf-hosted
Highlights
- Everything is a plugin: tools, model providers, services and UI modules stay decoupled around a thin kernel.
- Integrates TypeSafe Jev with Choice, Score and Noul primitives; decisions trigger automatically with no separate Jev screen.
- Optional Computer Use primitives (screenshot, mouse, keyboard, command execution) sit behind tool permissions and approvals.
- Quality gate as of 2026-10-07: ruff, mypy --strict on 125 files, 742 pytest cases and 57 frontend tests all green.
- MIT licensed; v1.0.0 developer preview built and tested mainly on Windows 10 with DeepSeek, per the README.
Quickstart
bash
git clone --depth 1 https://github.com/andyqiuqiubo/andy-harness.git
cd andy-harness/backend && uv sync --extra dev
cd ../frontend && pnpm install && pnpm buildWatch out
README and GitHub metadata reviewed 2026-10-17; the README is Chinese and long, states the project was developed on Windows 10 and has not been verified on macOS or Linux, and gives no benchmarks or pricing for the Jev path, which has no standalone interface.
Questions about this entry
Short answers drawn from the linked project or page.
It is a plugin-based open-source agent harness with a chat GUI, multi-provider model access (DeepSeek, Qwen, Doubao and custom OpenAI-compatible models), session and context management, an MCP client and marketplace, skills, scheduled tasks, long-term memory, sub-agent delegation, tracing and a Tauri desktop shell.
It integrates TypeSafe AI's Jev model with three primitives: Choice (option selection), Score (rating) and Noul (yes/no), so the agent can make structured judgments. The README notes Jev currently has only a plugin key configuration entry: decisions are triggered automatically by the model, with no independent call screen.
Clone the repo, run uv sync --extra dev in backend and pnpm install plus pnpm build in frontend, then double-click start-all.bat on Windows or run make backend and make frontend in two terminals. The front end lands at http://localhost:5173 and the backend health check at http://localhost:8000/api/health.
Python 3.11+, Node.js 22.12+ and pnpm 10+, with uv for Python packages; the make path also needs GNU Make. The README says the project was created and developed on Windows 10 and not verified on macOS or Linux, and the Tauri desktop shell additionally needs a Rust toolchain with rustc 1.77.2 or newer.
It supports DeepSeek, Qwen, Doubao and custom OpenAI-compatible providers, and the README says only DeepSeek's deepseek-v4-flash was deeply debugged. An OortCloud integration can claim a free 10 yuan model credit after a login authorization, and a Jev plugin key is needed for the Jev decision path.
Several backend capabilities have no UI entry: channels, Computer Use, Jev, checkpoint resume and user management are reachable only through the REST API, per the README. The frontend bundle is not code-split, Docker's frontend lacks a reverse proxy, and the project remains Windows-first with a DeepSeek focus.
The GitHub user andyqiuqiubo publishes it under the MIT license as a v1.0.0 developer preview, created 2026-09-24 and pushed 2026-10-16, with the README aimed at developers learning and extending agent systems.
Reactions & coverage
Posts, threads, and videos about this entry from around the web.
X: Jev in front of the HeyGen MCP
X: Playwright CLI + Jev vs Playwright MCP
X: A local MCP server that cuts computer use from 23s to 0.8s. Cartwright built Cortex, a local MCP server where…
X: A side panel that drives any site's WebMCP tools
More Jev resources like this
An event-driven multi-agent network where JEV observes live state and assigns pending work to available agents. Its AgentRouter reference implementation routes on every completion event instead of waiting for a round to end, while general-purpose agents with OpenAI Responses API tools execute work concurrently.
Jev Agents & MCP#community#github#agents
Community
Kelly-Leeee/AgentNetwork-jevPython
Kelly-Leeee/AgentNetwork-jevPythonA lightweight self-hosted System 1 decision daemon and MCP server for AI agents and multi-agent swarms: it evaluates state text against typed choice, score and noul questions in one forward pass, with Laya, Ollama and OpenAI-compatible backends.
Jev Agents & MCP#community#github#mcp
Community
axiomantic/local-systemonePython
axiomantic/local-systemonePythonSelf-hosted agent workspace platform from YaoApp that keeps all your agents and workspaces in one place across desktop, mobile, browser, or API, with Jev available inside it as TypeSafe AI's System One model for typed decisions with confidence.
Jev Agents & MCP#community#github#agents
Community
yaoapp/yaoGo
yaoapp/yaoGoChrome MV3 extension plus a local MCP server exposing 50 browser tools; its chrome_act_toward_goal micro-loop uses TypeSafe Jev System One or a local decider to pick the next action on an 85%+ pruned, Shadow-DOM-pierced DOM tree, dispatching trusted CDP events in logged-in Chrome without stealing focus.
Jev Agents & MCP#community#mcp#browser-automation
Community
GoldenLoaf24h/browserpawTypeScript
GoldenLoaf24h/browserpawTypeScriptAn open-source Jev-based MCP router: it discovers, catalogs, classifies and deduplicates every MCP server and tool you have, then exposes each agent only the small, authorized subset it needs for the task. Routing decisions are System One Choice/Score/Noul questions, scored by a pluggable engine at the edge or hosted.
Jev Agents & MCP#community#catalog-import#logicrw
Community
amantux/mcp-routerPython
amantux/mcp-routerPythonAn MCP server that lets Claude hand a whole browser task to Jev: Claude plans and verifies while Jev picks each click, type or select from a numbered table of visible controls in about 100 ms, using browser-use/jev-ultrafast's loop.
Jev Agents & MCP#community#catalog-import#madewithjev
Community
msalvalaggio/jev-reflexPython
msalvalaggio/jev-reflexPythonFrom 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.