User repository on GitHub
通用客服平台:ChatKit UI + Jev 决策 + Qwen 生成 + RAG 知识库 + 插件化 Tools + MCP 数据面,换行业只改 YAML 不改代码
- Forks
- 0
- Last pushed
- Updated Sep 2026
Repo stats from the GitHub API, cached Oct 2026.Homepage
Jev directory
CommunityJev Tools & CLIs
Jeves Desk
A general-purpose customer-service platform rebuilt from OpenAI's ChatKit sample: one Jev call returns intent, emotion, confidence, a recommended function and the right configured agent; a chat model drives MCP tools and built-in RAG answers from a knowledge base. Changing industry means editing YAML, not code.
- Stars
- 1
- Language
- JavaScript
- Reading time
- 3 min
- Published by
- Community
- Author
- dabaicai001
- Last verified
- Oct 2026
- Languages
- JavaScriptPython
- Capabilities
- Open sourceSelf-hosted
Highlights
- One Jev call per turn returns intent, emotion, confidence, a recommended function and which agent should take the case — Jev judges, it never writes the reply.
- Specialist agents (orders, products, knowledge base, after-sales, general) live in business.yaml, and Jev's dispatch shows live in the frontend panel.
- The chat model starts tool calls through an embedded MCP client, and code-side validation strips forged customer_id values before anything executes.
- Garbage intent gets a fixed friendly reply with no tool calls or generation; refunds, cancellations and tickets stay away from the model for humans.
- A local snapshot of the ChatKit UI assets (~7.4 MB) keeps the interface working without the CDN, and a pure-Python RAG answers through a query_knowledge tool.
Watch out
Grounded on the truncated Chinese-language README and GitHub metadata read on 2026-10-15; the excerpt ends inside the directory tree and does not show run commands, API keys or pricing, so none are quoted. The project derives from OpenAI's MIT-licensed openai-chatkit-advanced-samples with attribution retained, while GitHub metadata reports the license as NOASSERTION.
Questions about this entry
Short answers drawn from the linked project or page.
A general customer-service platform refactored from OpenAI's official ChatKit customer-support sample: the airline business is removed and abstracted so a change of industry means editing configuration rather than code. It bundles the ChatKit UI, a Python backend, Jev decisions, RAG and MCP tools.
ai/jev.py asks for intent, emotion, confidence, a recommended function and the agent to dispatch in one call — a structured decision rather than an answer. The frontend's right panel shows Jev's agent choice live through an agent_dispatch/update side effect.
They are configuration: the agents section of business.yaml defines each persona prompt and its available tools, covering orders, product advice, knowledge-base support, after-sales and a general agent. Adding or adjusting one is a YAML change, not a code change.
Change-type operations that need confirmation are deliberately not exposed to the model; they are handled by the human process. Handoff itself is executed by code (transfer_to_human) when the user asks, the emotion turns angry, or an unexposed change operation comes up.
Through an embedded MCP client — MCP is the data plane for customers, orders, products and tickets, so the platform never connects to business databases directly. The code side validates tool calls with pydantic, strips forged customer_id values and feeds results back to the model.
A built-in pure-Python vector retrieval, switchable to chroma, is exposed to the model as a query_knowledge tool. Tool results are fed back so the model continues on facts until it produces the final reply, and garbage-intent messages skip tools and generation entirely.
Yes: it is derived from openai/openai-chatkit-advanced-samples (MIT, Copyright 2025 OpenAI) and keeps the original license and attribution in LICENSE. GitHub metadata itself reports NOASSERTION, so read the bundled license before redistributing.
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
A local, open-source personal AI assistant for your job: mail, chats and tickets become tasks, agents you already use such as Claude Code, Codex and Gemini do the work, and nothing is sent until you approve. The current release is v0.3.8.3.
Jev Tools & CLIs#community#catalog-import#logicrw
Community
ldbumble/taskuaryPython
ldbumble/taskuaryPythonUse Jev from Claude Code, Codex CLI, Claude Desktop or any MCP client, paid for by free Netlify credits and locked with a key only you have: a key-locked proxy forwards through Netlify AI Gateway and an MCP server exposes jev_evaluate and jev_models.
Jev Agents & MCP#community#catalog-import#logicrw
Community
demircigoksu/jev-netlify-mcpPython
demircigoksu/jev-netlify-mcpPythonA desktop app for dictation, audio-file transcription and meeting notes on Windows, macOS and Linux, using local Whisper models or the OpenAI API, with an opt-in MCP server and local history API so agents can search saved transcripts.
Jev Tools & CLIs#community#catalog-import#logicrw
Community
knuckles92/openwhisperPython
knuckles92/openwhisperPythonA Chinese-language e-commerce customer-service demo on FastAPI, LangGraph and React: adaptive retrieval routing with hybrid BM25-plus-vector search, and intent classification switchable between TypeSafe Jev's structured choice interface and DeepSeek.
Jev Demos & Experiments#community#github#python
Community
gbh1234/Adaptive-ragPython
gbh1234/Adaptive-ragPythonE-commerce customer-service quality sandbox: Jev judges intent, emotion, urgency and risk per message, an 8-rule decision table routes to queues with SLAs, and dangerous agent phrasing is blocked with rewrite suggestions before it is sent.
Jev Demos & Experiments#community#catalog-import#logicrw
Community
kiler398/jev-demoTypeScript
kiler398/jev-demoTypeScriptToolkit, knowledge base and local dashboard for building with Jev via OpenRouter Alpha Decisions or the TypeSafe Direct API: a zero-dependency SQLite request log, a localhost REST server and Python CLI demos.
Jev Tools & CLIs#community#catalog-import#logicrw
Community
rileycarney/jevtoolsHTML
rileycarney/jevtoolsHTMLFrom 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.