Content Pilot / 内容领航员

AI-driven social media automation tool for Chinese platforms.

English | 中文

English

Features

Multi-platform support: Xiaohongshu, Douyin, Bilibili, Weibo
AI content generation: Qwen, GLM, OpenAI, Claude — with platform-optimized prompts
AI image cards: Auto-generate styled image cards (quote, title, list, minimal) from your content using AI
Web image search: Search and attach Unsplash images directly from the content editor
Web GUI: Browser-based dashboard for managing accounts, content, publishing, and schedules
Kanban view: Visual kanban board on the Publish page for managing drafts by status
Calendar view: Monthly calendar on the Schedule page showing upcoming scheduled posts
QR code login: Scan-to-login for all platforms with session persistence
Scheduled publishing: Cron-based scheduling with optimal posting times
Analytics tracking: Views, likes, comments, follower growth trends
Safety controls: Rate limiting, content validation, human review workflow
Anti-detection: Stealth browser, human-like delays, conservative publishing rates

Requirements

Python 3.11+
An API key: Qwen (阿里千问), GLM (智谱), OpenAI, or Anthropic (Claude)

Installation

git clone https://github.com/r04943083/content-pilot.git
cd content-pilot
bash install.sh

The install script will automatically:

Detect Python 3.11+ on your system
Install python3-venv if missing (requires sudo on Linux)
Create a .venv virtual environment
Install all Python dependencies
Install Playwright Chromium browser
Install Playwright system dependencies if missing (requires sudo on Debian/Ubuntu)
Create .env from template

When is sudo needed? The script only uses sudo for two things:

Installing the python3-venv package if your system Python is missing it
Running playwright install-deps chromium to install system libraries (Debian/Ubuntu only)

On macOS with Homebrew, sudo is generally not needed. You will be prompted before any sudo command runs.

Quick Start (GUI)

After installation, every time you open a new terminal:

cd content-pilot
source .venv/bin/activate
content-pilot gui

Then open http://127.0.0.1:8080 in your browser.

Step-by-Step Workflow

Configure API Key — Go to Settings page, select your AI provider and enter the API key
Login to Platforms — Go to Accounts page, click login button for each platform and scan QR code with your phone
Generate Content — Go to Content page, enter a topic, select platform and style, click Generate
Add Images — Use AI-generated image cards, search Unsplash for photos, or upload your own
Review & Edit — Preview the AI-generated content, edit if needed, then approve
Publish — Go to Publish page, select approved drafts and click Publish (use kanban view to see drafts by status)
Schedule (Optional) — Go to Schedule page to set up automated daily posting (use calendar view to visualize)

GUI Pages Overview

Page	Description
Dashboard	Overview of accounts and recent posts
Accounts	Scan QR code to login to each platform
Content	Generate AI content with image cards, web image search, and local upload
Publish	Manage and publish approved drafts; list view and kanban board
Schedule	Set up automated posting schedules with calendar overview
Settings	Configure AI provider and API keys

Content Styles

Style	Description
`tutorial`	Step-by-step guide
`review`	Product/service review
`lifestyle`	Personal life sharing
`knowledge`	Knowledge/educational
`story`	Narrative storytelling

Image Card Styles

When generating content, you can auto-create styled image cards:

Card Style	Description	Best For
`quote`	Large centered text with gradient background	Inspirational quotes, key takeaways
`title`	Title + subtitle layout	Cover images, thumbnails
`list`	Bullet-point list format	Tutorials, tips, checklists
`minimal`	Clean text-only design	Simple, elegant posts

Each platform has a default card style (e.g., Xiaohongshu defaults to quote, Bilibili to list).

Best Posting Times

Platform	Recommended Times
Xiaohongshu	7-9am, 12-1pm, 7-10pm
Douyin	12-1pm, 6-8pm, 9-11pm
Bilibili	11am-1pm, 5-7pm, 9-11pm
Weibo	8-9am, 12-1pm, 6-10pm

Configuration

Configuration uses three layers (later overrides earlier):

config/default.toml - Default settings
~/.content-pilot/config.toml - User settings
Environment variables with CP_ prefix (e.g., CP_AI__PROVIDER=openai)

You can configure everything via the Settings page in the GUI, or manually edit .env:

# Default provider is Qwen (千问):
CP_AI__QWEN_API_KEY=sk-...

# Or use GLM (智谱):
CP_AI__PROVIDER=glm
CP_AI__GLM_API_KEY=...

# Or OpenAI / Claude:
CP_AI__PROVIDER=openai   CP_AI__OPENAI_API_KEY=sk-...
CP_AI__PROVIDER=claude   CP_AI__ANTHROPIC_API_KEY=sk-ant-...

中文

功能特性

多平台支持: 小红书、抖音、B站、微博
AI 内容生成: 千问、智谱 GLM、OpenAI、Claude，针对各平台优化的 prompt
AI 图片卡片: 自动生成风格化配图卡片（引用、标题、清单、极简），从内容一键生成
网络图片搜索: 在内容编辑器中直接搜索 Unsplash 图片并添加到帖子
可视化界面: 基于浏览器的 Web GUI，管理账号、内容、发布和定时任务
看板视图: 发布页提供看板视图，按状态可视化管理草稿
日历视图: 定时任务页提供月历视图，直观查看排期
扫码登录: 所有平台支持扫码登录，会话持久化
定时发布: 基于 Cron 的定时调度，内置最佳发布时间建议
数据分析: 浏览、点赞、评论、粉丝增长趋势
安全控制: 频率限制、内容校验、人工审核流程
反检测: 隐身浏览器、仿人延迟、保守发布频率

系统要求

Python 3.11+
API Key: 千问 (Qwen)、智谱 (GLM)、OpenAI 或 Anthropic (Claude)

安装

git clone https://github.com/r04943083/content-pilot.git
cd content-pilot
bash install.sh

安装脚本会自动完成：

检测系统 Python 3.11+ 版本
缺少 python3-venv 时自动安装（Linux 下需要 sudo）
创建 .venv 虚拟环境
安装所有 Python 依赖
安装 Playwright Chromium 浏览器
缺少系统依赖时自动安装（Debian/Ubuntu 下需要 sudo）
从模板创建 .env 文件

何时需要 sudo？ 脚本仅在以下两种情况使用 sudo：

系统缺少 python3-venv 包时安装它
在 Debian/Ubuntu 上运行 playwright install-deps chromium 安装系统库

macOS + Homebrew 环境通常不需要 sudo。每次使用 sudo 前脚本会提示确认。

快速上手（GUI）

安装完成后，每次打开新终端需要先激活环境：

cd content-pilot
source .venv/bin/activate
content-pilot gui

然后在浏览器中打开 http://127.0.0.1:8080

操作流程

配置 API Key — 进入 Settings 页面，选择 AI 提供商并输入 API Key
登录平台 — 进入 Accounts 页面，点击各平台的登录按钮，用手机扫码
生成内容 — 进入 Content 页面，输入主题，选择平台和风格，点击 Generate
添加配图 — 使用 AI 生成图片卡片、搜索 Unsplash 图片，或上传本地图片
审核编辑 — 预览 AI 生成的内容，按需编辑，然后批准
发布 — 进入 Publish 页面，选择已批准的草稿，点击 Publish（可切换看板视图按状态管理）
定时任务（可选） — 进入 Schedule 页面设置每日自动发帖（可切换日历视图查看排期）

GUI 页面说明

页面	功能
Dashboard	账号和最近帖子概览
Accounts	各平台扫码登录
Content	AI 内容生成、图片卡片、网络图搜索、本地上传
Publish	管理和发布草稿，支持列表和看板两种视图
Schedule	设置定时发布任务，带日历概览
Settings	配置 AI 提供商和 API Keys

内容风格

风格	说明
`tutorial`	教程 / 步骤讲解
`review`	测评 / 产品点评
`lifestyle`	生活分享
`knowledge`	知识科普
`story`	故事叙述

图片卡片风格

生成内容时可自动创建风格化配图卡片：

卡片风格	说明	适用场景
`quote`	大字体居中，渐变背景	名言金句、核心观点
`title`	标题+副标题布局	封面图、缩略图
`list`	要点列表形式	教程、技巧、清单
`minimal`	纯文字，干净设计	简约风格帖子

每个平台有默认卡片风格（如小红书默认 quote，B站默认 list）。

最佳发布时间

平台	推荐时段
小红书	7-9点、12-13点、19-22点
抖音	12-13点、18-20点、21-23点
B站	11-13点、17-19点、21-23点
微博	8-9点、12-13点、18-22点

配置

配置使用三层结构（后者覆盖前者）：

config/default.toml - 默认设置
~/.content-pilot/config.toml - 用户设置
环境变量，前缀为 CP_（如 CP_AI__PROVIDER=openai）

你可以在 GUI 的 Settings 页面配置一切，或者手动编辑 .env：

# 默认使用千问 (Qwen):
CP_AI__QWEN_API_KEY=sk-...

# 或使用智谱 GLM:
CP_AI__PROVIDER=glm
CP_AI__GLM_API_KEY=...

# 或使用 OpenAI / Claude:
CP_AI__PROVIDER=openai   CP_AI__OPENAI_API_KEY=sk-...
CP_AI__PROVIDER=claude   CP_AI__ANTHROPIC_API_KEY=sk-ant-...

CLI Reference (Advanced)

For users who prefer command-line interface:

# Login via QR code
content-pilot login --platform xiaohongshu

# Check account status
content-pilot status

# Generate content
content-pilot generate --topic "Python学习技巧" --platform xiaohongshu --style tutorial

# Publish (dry-run first)
content-pilot publish --content-id 1 --dry-run
content-pilot publish --content-id 1

# Analytics
content-pilot analytics summary

# Schedule management
content-pilot schedule add --name "每日发帖" --platform xiaohongshu --topic "Python技巧" --cron "0 20 * * *"
content-pilot schedule list
content-pilot run   # start daemon

Docker Deployment

# First login locally (requires QR scan)
content-pilot login --platform xiaohongshu

# Then run daemon with Docker
docker compose up -d

Development

Running Tests

source .venv/bin/activate
pytest                    # run all tests
pytest tests/ -v          # verbose output
pytest --cov              # with coverage report

Linting

ruff check src/ tests/    # check for lint errors
ruff format src/ tests/   # auto-format code

Project Structure

src/content_pilot/
  cli.py              # CLI entry point
  app.py              # Core application logic
  config.py           # Settings and configuration
  content/            # Content generation and card templates
  gui/                # NiceGUI web interface
    pages/            # Dashboard, Accounts, Content, Publish, Schedule, Settings
    components/       # Reusable UI components (image picker, nav, cards)
    i18n/             # Internationalization (English + Chinese)
tests/                # pytest test suite
config/               # Default configuration files

Troubleshooting

Browser / Playwright Issues

Chromium fails to launch with missing library errors (libnspr4.so, libnss3.so, etc.)

This means Playwright system dependencies are not installed. Run:

sudo playwright install-deps chromium

install.sh fails at the Playwright step

If you are on a non-Debian system (Fedora, Arch, macOS), the script may not auto-install system deps. Install Chromium dependencies for your distro manually, or try:

playwright install chromium

and check the error output for missing libraries.

Browser launches but pages time out or fail to load

Make sure you are not behind a corporate proxy that blocks Chromium network access. If using WSL2, ensure your network is working (curl https://www.baidu.com).

Installation Issues

python3-venv install fails

On Ubuntu/Debian, make sure you have the correct version:

sudo apt install python3.12-venv   # replace 3.12 with your Python version

pip install fails with build errors

Ensure you have build essentials:

sudo apt install build-essential python3-dev   # Debian/Ubuntu

GUI Issues

GUI starts but page is blank or shows errors

Try clearing your browser cache or opening in an incognito window. The GUI runs on NiceGUI which uses WebSocket connections -- make sure nothing is blocking port 8080.

Port 8080 is already in use

Another process is using the port. Find and stop it, or specify a different port:

content-pilot gui --port 8081

FAQ

Q: bash install.sh shows permission error? The script needs sudo only for installing python3-venv (if missing) and Playwright system dependencies (Debian/Ubuntu). It will prompt before running any sudo command.

Q: source .venv/bin/activate reports No such file? Delete the old .venv and run bash install.sh again:

rm -rf .venv
bash install.sh

Q: Login reports TargetClosedError / libnspr4.so not found? Chromium is missing system dependencies. Install them:

sudo playwright install-deps chromium

Q: How to switch AI provider? Edit .env and set CP_AI__PROVIDER=qwen|glm|openai|claude with the corresponding API key. Default is qwen (千问), recommended for users in China.

Q: How do I add images to my posts? On the Content page, after generating text, use the image picker tabs: "Code-Generated Cards" creates AI-designed image cards, "Web Search" finds photos from Unsplash, and "Upload" lets you use local files.

Q: What is the kanban view on the Publish page? Click the kanban icon in the toolbar to switch from list view to a visual board that groups your drafts by status (draft, approved, published, etc.).

Q: Can I use this on macOS? Yes. Install Python 3.11+ via Homebrew (brew install python@3.12), then run bash install.sh. No sudo should be needed on macOS.

Legal Disclaimer

This tool is for learning and research purposes only. Users must comply with the terms of service of each platform and relevant laws and regulations. According to China's "Provisions on the Administration of Deep Synthesis of Internet Information Services", AI-generated content should be labeled. Please comply with relevant regulations when using.

License

MIT

Name		Name	Last commit message	Last commit date
config		config
src/content_pilot		src/content_pilot
tests		tests
.env.example		.env.example
.gitignore		.gitignore
Dockerfile		Dockerfile
LICENSE		LICENSE
Makefile		Makefile
README.md		README.md
docker-compose.yml		docker-compose.yml
install.sh		install.sh
pyproject.toml		pyproject.toml

Latest commit

History

Folders and files

Repository files navigation