Skip to main content
QvQChat

QvQChat

让 AI 像真人一样参与群聊。

PyPI Python Docker License ErisPulse


基于 ErisPulse 的智能对话模块。多模型池 + 行为绑定 + 多智能体人格 + 长期记忆 + 知识库 + MCP 工具,配合全功能 Dashboard 管理面板——适配器、模型、行为配置全部可视化完成。

核心特性

👀 会窥屏

群聊中默默观察,只在被 @、有人叫名字或话题相关时才开口。预测模式(低 token)批量判断是否参与,不抢话、不刷屏

🕐 有生活感

打字延迟随回复长短变化、清晨迷糊深夜放得开、错字后自己纠正、偶尔半句发出——像真人一样不完美

💬 主动搭话

冲动值驱动:聊天越热闹越想开口;睡觉时段不打扰、没人理就冷却;没有真正想说的就保持沉默

🧠 模型池 + 行为绑定

任意 OpenAI 兼容 API 汇入模型池,按能力标记(chat/vision/tools),多模型冗余故障自动切换,行为级独立分配

📝 会记忆

回复后自动提取长期记忆,群聊支持混合/仅发送者两种模式,下次聊天自然提起

🖥️ 全功能 Dashboard

21 个 API 的 Web 管理面板:模型、行为、智能体、知识库、表情包、配置即时生效,无需手动编辑文件


工作原理

一条消息从进入到回复的完整处理链:

graph LR
    subgraph Platforms[平台]
        P["QQ / 云湖 / OneBot / ..."]
    end

    Adapter["ErisPulse 适配器层<br/>OB12 标准事件"]

    subgraph QvQChat
        Gate["回复判定<br/>@感知 → 活跃模式 → 预测/窥屏策略"]
        Pipeline["提示词注入管线<br/>身份 → 规则 → 场景 → 知识库 → 时间 → 情绪"]
        Dialogue["对话行为<br/>模型池故障转移 + MCP 工具循环"]
        Humanize["拟人化后处理<br/>错字纠正 / 半句发出 / 打字延迟"]
        Send["消息发送<br/>表情包标签 / 多条分段 / 语音"]
    end

    Memory["记忆子系统<br/>短期历史 + 长期提取"]
    Session["会话管理<br/>冲动值 + 速率限制"]
    Proactive["主动发起<br/>冲动值门槛检查"]

    P -->|"OB12 事件"| Adapter
    Adapter --> Gate
    Gate -->|"回复"| Pipeline
    Pipeline --> Dialogue
    Dialogue --> Humanize
    Humanize --> Send
    Send -.->|"回复"| P
    Adapter -.->|"累积"| Memory
    Adapter -.->|"冲动值 +"| Session
    Session -.->|"门槛满足"| Proactive
    Proactive -.->|"主动开口"| Dialogue

Dashboard

所有配置一站式完成,修改即时生效。

Dashboard 概览

行为管理 — 自由定制 AI 人格

内置 7 种行为 + 自定义行为,独立配置提示词、模型、温度。

行为管理

基础设置 + 表情包

窥屏/速率/拟人化参数一目了然;自定义表情包由 AI 按场景自主发送。

基础设置

表情包


功能一览

模块 说明
🧠 模型池 多模型管理,能力标记(对话/视觉/工具调用),故障自动切换
🎭 行为系统 自定义行为、独立提示词、模型分配、触发模式(始终/预测)
👥 多智能体 猫娘/傲娇/温柔大姐姐等人格模板,按群/用户绑定
👀 窥屏模式 群聊默认观察,被 @ / 叫名字 / 活跃模式时才回复
🔮 预测模式 低 token:累积 N 条消息批量判断,命中触发词才进入对话
🕐 拟人化 打字延迟、时间感知、情绪感知、错字纠正、半句发出、多条消息
💬 主动搭话 冲动值驱动,睡眠不打扰、没人理就冷却、宁可沉默不尬聊
📝 记忆系统 自动提取长期记忆,群聊支持混合/仅发送者两种模式
📚 知识库 文档注入对话上下文,支持分类、标签、自动搜索
🔧 MCP 工具 函数调用,让 AI 调用外部 API
🎙️ 语音合成 <|voice style="语气"|> 语音标签,CosyVoice2 合成
🖥️ Dashboard 全功能 Web 面板,所有配置即时生效

快速开始

Docker(推荐)

docker run -d --name qvqchat -p 8000:8000 --restart unless-stopped ghcr.io/wsu2059q/erispulse-qvqchat:latest

启动后打开 http://localhost:8000/Dashboard,适配器、AI 模型、行为配置全部在面板中完成。

使用 pip 安装
pip install ErisPulse
pip install ErisPulse-QvQChat
ep run

初始配置

  1. Dashboard 中添加适配器(如 YunhuAdapter、OneBot11Adapter)
  2. 添加 AI 模型(OpenAI / DeepSeek / SiliconFlow 等任意兼容 API),标记能力
  3. 在行为管理中为各行为分配模型——完成,开始聊天

文档


许可证

本项目基于 MIT License 开源。

驱动于 ErisPulse —— 一次编写,部署到多个聊天平台。

Metadata

Release files for ErisPulse-QvQChat 3.2.1

For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.

Source distribution (sdist)

Source distribution for ErisPulse-QvQChat 3.2.1
File Size Uploaded
erispulse_qvqchat-3.2.1.tar.gz 152.7 kB Details

Built distribution (wheel)

Table of built distributions (wheels) for ErisPulse-QvQChat 3.2.1
File Interpreter ABI Platform
erispulse_qvqchat-3.2.1-py3-none-any.whl Python 3 none any Details

Total release size: 321.4 kB

Release files / erispulse_qvqchat-3.2.1.tar.gz

Download URL erispulse_qvqchat-3.2.1.tar.gz
Size 152.7 kB
Tags Source
SHA-256 checksum
How to use checksums
3a8054bd4874cc05e9d3e835fc87093b0dc476fa58184c53a50e08a0ae5d0dbf
BLAKE2b-256 checksum
How to use checksums
6ea9bb0399706c885e8538b0d3d797187520a0e773e2bda0a709fd0d56cbad66
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
Yes
Uploaded via twine/7.0.0 CPython/3.13.14

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Sep 10, 2026.

Transparency log

Release files / erispulse_qvqchat-3.2.1-py3-none-any.whl

Download URL erispulse_qvqchat-3.2.1-py3-none-any.whl
Size 168.7 kB
Tags Python 3
SHA-256 checksum
How to use checksums
e5ead3ce83e091bd4fd9bd6bd46c6c402d5ed675f10a87610a7f518973445621
BLAKE2b-256 checksum
How to use checksums
86c7f907e434c0e83461ea0b6b9fe2f8439c1eab9a74572f4c5f6d992bd931f3
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
Yes
Uploaded via twine/7.0.0 CPython/3.13.14

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Sep 10, 2026.

Transparency log

Release history Release notifications | RSS feed

This release

3.2.1 This release

2 release files

3.2.0

2 release files

3.0.0

2 release files

2.7.0

2 release files

2.6.4

2 release files

2.6.2

2 release files

2.6.1

2 release files

2.6.0

2 release files

2.5.0

2 release files

2.3.1

2 release files

2.3.0

2 release files

2.2.1

2 release files

2.2.0

2 release files

2.1.0

2 release files

2.0.0

2 release files

Anthropic, PBC Visionary sponsor Bloomberg Visionary sponsor Hudson River Trading Visionary sponsor Meta Visionary sponsor NVIDIA Visionary sponsor Microsoft Sustainability sponsor Depot Continuous Integration AWS Cloud computing and Security Sponsor Datadog Monitoring Fastly CDN Google Download Analytics Sentry Error logging StatusPage Status page