aiticker

Pack · #362 trending skills repo · ↗ +1 stars this week

258-Canvas

ken571571/258-Canvas

AI-powered infinite canvas — aggregate multiple AI platforms for image & video generation, LLM chat, and ComfyUI workflow execution. Built-in Agent system with knowledge bases, custom Python skills, and encrypted distribution. Connect your local ComfyUI backend for seamless workflow automation. Zero-config LAN collaboration with bilingual UI.

Frontend & UIBackend & APIsAgents & orchestrationMedia & games
Install ↓GitHub ↗
Stars
59
This week
+1
Forks
0
Open issues
0
Latest release
v2.5.61
Star trend · last 9 days+2 ★
Oct 29, 2026: 57 starsOct 30, 2026: 58 starsOct 31, 2026: 58 starsNov 1, 2026: 58 starsNov 2, 2026: 58 starsNov 3, 2026: 58 starsNov 4, 2026: 58 starsNov 5, 2026: 58 starsNov 6, 2026: 59 starsNov 7, 2026: 59 stars

Get started

Install the skills

npx skills add ken571571/258-Canvas

Installs into Claude Code, Codex, Cursor, OpenCode and the other agents it finds on your machine.

Documentation

README

View on GitHub ↗
🇺🇸 English (click to switch language)

∞ 258 Canvas

AI-powered infinite canvas — aggregate multiple AI platforms for image & video generation, LLM chat, and ComfyUI workflow execution. Built-in Agent system with knowledge bases, custom Python skills, and encrypted distribution. Connect your local ComfyUI backend for seamless workflow automation. Zero-config LAN collaboration with bilingual UI.

Version Python License

✨ Features

  • 🎨 Infinite Canvas — visual node editor, drag-and-drop AI workflows
  • 🤖 Multi-Platform — OpenAI / DeepSeek / Gemini / Volcengine / ModelScope / RunningHub / APIMart
  • 🖼️ Image Generation — text-to-image, image-to-image, Loop batch pipelines
  • 🎬 Video Generation — Veo3 and other video models
  • 💬 LLM Chat — streaming/non-streaming, multi-conversation
  • 🧠 Agent System — ReAct execution engine, knowledge bases, custom Python skills
  • ⚡ ComfyUI Integration — workflow management + async execution, local & LAN backends with one-click LAN auto-discovery
  • 🔒 Encrypted Agent — AES-256-GCM, machine-bound, export/distribution ready
  • 🌐 LAN Collaboration — zero-config multi-device, optimistic locking
  • 🌍 Bilingual UI — 483 translation keys, 100% Chinese/English coverage

📦 Requirements

  • Python 3.10+
  • Windows / macOS / Linux
  • No database or external services required

🚀 Quick Start

# 1. Clone
git clone https://github.com/ken571571/258-Canvas.git
cd 258-Canvas

# 2. Install dependencies
pip install -r requirements.txt

# 3. (Optional) Configure API keys
cp API/.env.example API/.env
# Edit API/.env with your keys

# 4. Run
python run.py        # macOS / Linux
# or double-click 启动服务.bat on Windows
# or bash mac启动服务.sh on macOS (auto-setup + auto-open browser)

# 5. Open browser → http://127.0.0.1:3571
  • Localhost access: no password required
  • LAN access: enter preset password (default 258)

🖥️ Using ComfyUI on Another LAN Computer

Run ComfyUI on a powerful PC and create from any other machine on the same network — no ComfyUI install needed on the client.

On the ComfyUI computer (one-time setup):

  1. Make ComfyUI accept LAN requests: add --listen 0.0.0.0 to its launch command:
# Windows (all-in-one pack)
.\python_embeded\python.exe -s ComfyUI\main.py --windows-standalone-build --listen 0.0.0.0
# macOS / Linux
python3 main.py --listen 0.0.0.0
  1. Open the firewall (Windows, admin PowerShell; Linux: sudo ufw allow 8188/tcp; macOS needs no setup):
netsh advfirewall firewall add rule name="ComfyUI 8188" dir=in action=allow protocol=tcp localport=8188

On the Infinite Canvas computer:

Open Settings → ComfyUI Backends, click 🔍 Scan LAN to Auto-discover, then add the discovered backend (or enter IP:8188 manually). ComfyUI nodes work immediately.

⚠ When enabled, every device on the same LAN can access that ComfyUI. Do not enable it on public WiFi.

📖 Project Structure

├── server/         # Python FastAPI backend
├── static/         # Frontend (18 JS modules, 9 pages)
├── tests/          # Tests (241 cases)
├── workflows/      # ComfyUI workflow templates
├── agents/         # Agent data
├── canvases/       # Canvas data
├── skills/         # Custom skills
└── scripts/        # Test scripts

🧪 Run Tests

RATE_LIMIT_ENABLED=0 python -m unittest discover -s tests

⚠️ License

Source Available — free for commercial use, study, and internal use. Redistribution as a commercial product is prohibited unless explicitly authorized.

See LICENSE.

🇨🇳 中文（点击切换语言）

∞ 258 无限画布

AI 驱动的无限创作平台 — 聚合多 AI 平台的画布式创作工具。内置 Agent 智能体系统，支持知识库、自定义 Python 技能和加密分发。可对接本地 ComfyUI 后端实现无缝工作流自动化。零配置局域网协作，中英双语界面。

Version Python License

✨ 功能

  • 🎨 无限画布 — 可视化节点编辑器，拖拽连线构建 AI 工作流
  • 🤖 多平台聚合 — OpenAI / DeepSeek / Gemini / 火山方舟 / ModelScope / RunningHub / APIMart
  • 🖼️ 文生图 / 图生图 — 多模型图片生成，支持 Loop 批量管线
  • 🎬 视频生成 — 支持 Veo3 等视频模型
  • 💬 LLM 对话 — 流式/非流式聊天，多对话管理
  • 🧠 Agent 智能体 — ReAct 执行引擎，支持知识库和自定义 Python 技能
  • ⚡ ComfyUI 集成 — 工作流管理 + 异步执行，本地/局域网后端 + 局域网一键自动发现
  • 🔒 加密 Agent — AES-256-GCM 加密，机器绑定，支持付费分发
  • 🌐 局域网协作 — 零配置多端使用，乐观锁防冲突
  • 🌍 中英双语 — 483 个翻译键，100% 覆盖

📦 环境要求

  • Python 3.10+
  • Windows / macOS / Linux
  • 无需数据库，无需额外服务

🚀 快速启动

# 1. 克隆
git clone https://github.com/ken571571/258-Canvas.git
cd 258-Canvas

# 2. 安装依赖
pip install -r requirements.txt

# 3. （可选）配置 API Key
cp API/.env.example API/.env
# 编辑 API/.env，填入你的密钥

# 4. 运行
python run.py        # macOS / Linux
# Windows 双击 启动服务.bat
# macOS 运行 bash mac启动服务.sh（自动安装依赖 + 自动打开浏览器）

# 5. 打开浏览器 → http://127.0.0.1:3571
  • 本机免密访问
  • 局域网输入预设密码（默认 258）

🖥️ 局域网调用另一台电脑的 ComfyUI

让高性能电脑运行 ComfyUI，同一局域网内的其他电脑负责创作——客户端无需安装 ComfyUI。

在 ComfyUI 电脑上（一次性设置）：

  1. 让 ComfyUI 接受局域网请求：在启动命令末尾加上 --listen 0.0.0.0：
# Windows 整合包
.\python_embeded\python.exe -s ComfyUI\main.py --windows-standalone-build --listen 0.0.0.0
# macOS / Linux
python3 main.py --listen 0.0.0.0
  1. 放行防火墙（Windows 以管理员身份打开 PowerShell 执行；Linux：sudo ufw allow 8188/tcp；macOS 默认无需配置）：
netsh advfirewall firewall add rule name="ComfyUI 8188" dir=in action=allow protocol=tcp localport=8188

在无限画布电脑上：

打开 设置 → ComfyUI 后端，点击 🔍 扫描局域网自动发现，添加发现的后端（也可手动填写 IP:8188），ComfyUI 节点立即可用。

⚠ 开启后同一局域网内的所有设备都能访问该 ComfyUI，请不要在公共 WiFi 下开启。

📖 项目结构

├── server/         # Python FastAPI 后端
├── static/         # 前端（18 个 JS 模块 / 9 个页面）
├── tests/          # 测试（241 个用例）
├── workflows/      # ComfyUI 工作流模板
├── agents/         # Agent 数据
├── canvases/       # 画布数据
├── skills/         # 自定义技能
└── scripts/        # 测试脚本

🧪 运行测试

RATE_LIMIT_ENABLED=0 python -m unittest discover -s tests

⚠️ 许可

源码可用（Source Available） — 允许免费商用、学习研究和内部使用。禁止将源码封装为商业产品出售，除非获得明确授权。

详见 LICENSE。

Made with ❤️ by ken571571

Releases

Recent versions

All releases ↗
  1. v2.5.61