LLM Daily - Telegram Knowledge Base CLI

输入链接，自动解析内容，生成双语摘要和标签，一键发布到 Telegram 频道

Input a link, auto-parse content, generate bilingual summaries and tags, publish to Telegram with one click

Language / 语言: 中文 | English

中文

输入链接 → 自动解析内容 → 生成双语摘要和标签 → 一键发布到 Telegram 频道

输出示例：

输入: https://arxiv.org/abs/2312.xxxxx


输出到 Telegram:
📌 稳定大语言模型强化学习：公式化方法与实践

📅 2024-12-01

📝 本文提出了一种新颖的大语言模型强化学习公式...

🔗 https://arxiv.org/abs/2312.xxxxx

🏷️ #论文 #Paper #大语言模型 #LLM #强化学习 #RL

────────────────────

📌 Stabilizing Reinforcement Learning with LLMs: Formulation and Practices

📅 2024-12-01

📝 This paper proposes a novel formulation for RL with LLMs...

🔗 https://arxiv.org/abs/2312.xxxxx

🏷️ #Paper #paper #LLM #llm #RL #rl

支持的链接类型

类型	状态	说明
ArXiv	✅ 已支持	自动提取论文标题、摘要、发布日期
微信公众号	✅ 已支持	解析文章内容
通用网页	✅ 已支持	自动提取正文内容
知乎	⚠️ 受限	有反爬保护，建议手动复制内容使用文本模式
GitHub	🚧 计划中	解析 README、仓库信息
HuggingFace	🚧 计划中	Papers、Models、Datasets
小红书	🚧 计划中	笔记内容
Twitter/X	🚧 计划中	推文内容

核心功能

自动解析: 输入链接，自动抓取网页内容
双语输出: LLM 生成中英文标题和摘要
智能标签: 从预设标签匹配 + 自动生成新标签
日期提取: 自动提取文章发布日期
一键发布: 确认后直接发布到 Telegram 频道

安装

git clone https://github.com/0xPabloxx/telegram_knowledge_base_agent.git
cd telegram_knowledge_base_agent

python3 -m venv venv
source venv/bin/activate

pip install -e .

配置

创建 .env 文件：

# Telegram
KB_TELEGRAM_BOT_TOKEN=your_bot_token
KB_TELEGRAM_CHANNEL_ID=@your_channel

# LLM (支持: deepseek, openai, gemini, anthropic, kimi, minimax, glm)
KB_LLM_PROVIDER=deepseek
DEEPSEEK_API_KEY=your_api_key

# 预设标签
KB_PRESET_TAGS=Paper,LLM,Agent,Research,Tutorial

使用

CLI 模式（电脑端）

source venv/bin/activate
kb

然后输入链接即可：

▶ 有什么想收藏的？
https://arxiv.org/abs/2312.xxxxx

Bot 模式（手机端）

通过 Telegram Bot 在手机上发送链接，自动解析并发布到频道。

1. 额外配置

在 .env 中添加：

# 你的 Telegram User ID（在 Telegram 搜索 @userinfobot 获取）
KB_BOT_ADMIN_USER_ID=123456789

# 发布模式: false=预览确认, true=直接发布
KB_BOT_AUTO_PUBLISH=false

2. 启动 Bot

source venv/bin/activate
kb-serve

3. 使用方式

在手机上打开你的 Bot，发送链接：

Bot 自动解析并生成双语摘要
点击按钮选择标签
点击「发布」发送到频道

Bot 命令：

/start - 开始使用
/mode - 切换「直接发布」/「预览确认」模式
/cancel - 取消当前操作
/help - 帮助信息

4. 持久化运行

方法 A：后台运行（简单）

nohup kb-serve > ~/kb-bot.log 2>&1 &

查看日志：tail -f ~/kb-bot.log

停止服务：pkill -f kb-serve

方法 B：macOS 开机自启（推荐）

创建 launchd 配置文件：

cat > ~/Library/LaunchAgents/com.kb.bot.plist << 'EOF'
<?xml version="1.0" encoding="UTF-8"?>
<!DOCTYPE plist PUBLIC "-//Apple//DTD PLIST 1.0//EN" "http://www.apple.com/DTDs/PropertyList-1.0.dtd">
<plist version="1.0">
<dict>
    <key>Label</key>
    <string>com.kb.bot</string>
    <key>ProgramArguments</key>
    <array>
        <string>/bin/bash</string>
        <string>-c</string>
        <string>source /Users/YOUR_USERNAME/Projects/telegram_channel_bot/venv/bin/activate && kb-serve</string>
    </array>
    <key>WorkingDirectory</key>
    <string>/Users/YOUR_USERNAME/Projects/telegram_channel_bot</string>
    <key>RunAtLoad</key>
    <true/>
    <key>KeepAlive</key>
    <true/>
    <key>StandardOutPath</key>
    <string>/Users/YOUR_USERNAME/Library/Logs/kb-bot.log</string>
    <key>StandardErrorPath</key>
    <string>/Users/YOUR_USERNAME/Library/Logs/kb-bot.error.log</string>
    <key>EnvironmentVariables</key>
    <dict>
        <key>PATH</key>
        <string>/usr/local/bin:/usr/bin:/bin</string>
    </dict>
</dict>
</plist>
EOF

注意：将 YOUR_USERNAME 替换为你的用户名

加载服务：

launchctl load ~/Library/LaunchAgents/com.kb.bot.plist

常用命令：

# 启动
launchctl start com.kb.bot

# 停止
launchctl stop com.kb.bot

# 卸载（停止并移除自启）
launchctl unload ~/Library/LaunchAgents/com.kb.bot.plist

# 查看日志
tail -f ~/Library/Logs/kb-bot.log

English

Input a link → Auto-parse content → Generate bilingual summaries & tags → Publish to Telegram

Output Example:

Input: https://arxiv.org/abs/2312.xxxxx

Output to Telegram:
📌 Chinese Title (auto-translated)

📅 2024-12-01

📝 Chinese summary...

🔗 https://arxiv.org/abs/2312.xxxxx

🏷️ #Chinese #Tags

────────────────────

📌 English Title

📅 2024-12-01

📝 English summary...

🔗 https://arxiv.org/abs/2312.xxxxx

🏷️ #English #Tags

Supported Link Types

Type	Status	Description
ArXiv	✅ Supported	Auto-extract paper title, abstract, date
WeChat Articles	✅ Supported	Parse article content
General Web	✅ Supported	Auto-extract main content
Zhihu	⚠️ Limited	Anti-scraping protection, use text mode instead
GitHub	🚧 Planned	Parse README, repo info
HuggingFace	🚧 Planned	Papers, Models, Datasets
Xiaohongshu	🚧 Planned	Note content
Twitter/X	🚧 Planned	Tweet content

Core Features

Auto-parsing: Input a link, automatically scrape web content
Bilingual Output: LLM generates Chinese and English titles/summaries
Smart Tags: Match from presets + auto-generate new tags
Date Extraction: Auto-extract article publish date
One-click Publish: Publish to Telegram channel after confirmation

Installation

git clone https://github.com/0xPabloxx/telegram_knowledge_base_agent.git
cd telegram_knowledge_base_agent

python3 -m venv venv
source venv/bin/activate

pip install -e .

Configuration

Create .env file:

# Telegram
KB_TELEGRAM_BOT_TOKEN=your_bot_token
KB_TELEGRAM_CHANNEL_ID=@your_channel

# LLM (supports: deepseek, openai, gemini, anthropic, kimi, minimax, glm)
KB_LLM_PROVIDER=deepseek
DEEPSEEK_API_KEY=your_api_key

# Preset Tags
KB_PRESET_TAGS=Paper,LLM,Agent,Research,Tutorial

Usage

CLI Mode (Desktop)

source venv/bin/activate
kb

Then input a link:

▶ What do you want to save?
https://arxiv.org/abs/2312.xxxxx

Bot Mode (Mobile)

Use Telegram Bot on your phone to send links, auto-parse and publish to channel.

1. Additional Configuration

Add to .env:

# Your Telegram User ID (search @userinfobot on Telegram to get it)
KB_BOT_ADMIN_USER_ID=123456789

# Publish mode: false=preview & confirm, true=auto publish
KB_BOT_AUTO_PUBLISH=false

2. Start Bot

source venv/bin/activate
kb-serve

3. How to Use

Open your Bot on phone, send a link:

Bot auto-parses and generates bilingual summary
Click buttons to select tags
Click "Publish" to send to channel

Bot Commands:

/start - Get started
/mode - Toggle "auto publish" / "preview & confirm" mode
/cancel - Cancel current operation
/help - Help info

4. Keep Running

Method A: Background Run (Simple)

nohup kb-serve > ~/kb-bot.log 2>&1 &

View logs: tail -f ~/kb-bot.log

Stop service: pkill -f kb-serve

Method B: macOS Auto-start (Recommended)

Create launchd config file:

cat > ~/Library/LaunchAgents/com.kb.bot.plist << 'EOF'
<?xml version="1.0" encoding="UTF-8"?>
<!DOCTYPE plist PUBLIC "-//Apple//DTD PLIST 1.0//EN" "http://www.apple.com/DTDs/PropertyList-1.0.dtd">
<plist version="1.0">
<dict>
    <key>Label</key>
    <string>com.kb.bot</string>
    <key>ProgramArguments</key>
    <array>
        <string>/bin/bash</string>
        <string>-c</string>
        <string>source /Users/YOUR_USERNAME/Projects/telegram_channel_bot/venv/bin/activate && kb-serve</string>
    </array>
    <key>WorkingDirectory</key>
    <string>/Users/YOUR_USERNAME/Projects/telegram_channel_bot</string>
    <key>RunAtLoad</key>
    <true/>
    <key>KeepAlive</key>
    <true/>
    <key>StandardOutPath</key>
    <string>/Users/YOUR_USERNAME/Library/Logs/kb-bot.log</string>
    <key>StandardErrorPath</key>
    <string>/Users/YOUR_USERNAME/Library/Logs/kb-bot.error.log</string>
    <key>EnvironmentVariables</key>
    <dict>
        <key>PATH</key>
        <string>/usr/local/bin:/usr/bin:/bin</string>
    </dict>
</dict>
</plist>
EOF

Note: Replace YOUR_USERNAME with your actual username

Load the service:

launchctl load ~/Library/LaunchAgents/com.kb.bot.plist

Common commands:

# Start
launchctl start com.kb.bot

# Stop
launchctl stop com.kb.bot

# Unload (stop and remove auto-start)
launchctl unload ~/Library/LaunchAgents/com.kb.bot.plist

# View logs
tail -f ~/Library/Logs/kb-bot.log

Supported LLM Providers

Provider	Model
DeepSeek	deepseek-chat
OpenAI	gpt-4o-mini
Anthropic	claude-3-5-sonnet
Google Gemini	gemini-1.5-flash
Kimi	moonshot-v1-8k
MiniMax	abab6.5s-chat
GLM	glm-4-flash

License

MIT

Name		Name	Last commit message	Last commit date
Latest commit History 11 Commits
kb		kb
.env.example		.env.example
.gitignore		.gitignore
README.md		README.md
TODO.md		TODO.md
demo.gif		demo.gif
pyproject.toml		pyproject.toml

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Repository files navigation

LLM Daily - Telegram Knowledge Base CLI

中文

支持的链接类型

核心功能

安装

配置

使用

CLI 模式（电脑端）

Bot 模式（手机端）

1. 额外配置

2. 启动 Bot

3. 使用方式

4. 持久化运行

English

Supported Link Types

Core Features

Installation

Configuration

Usage

CLI Mode (Desktop)

Bot Mode (Mobile)

1. Additional Configuration

2. Start Bot

3. How to Use

4. Keep Running

Supported LLM Providers

License

About

Uh oh!

Releases

Packages

Contributors 2

Uh oh!

Languages

0xPabloxx/telegram_knowledge_base_agent

Folders and files

Latest commit

History

Repository files navigation

LLM Daily - Telegram Knowledge Base CLI

中文

支持的链接类型

核心功能

安装

配置

使用

CLI 模式（电脑端）

Bot 模式（手机端）

1. 额外配置

2. 启动 Bot

3. 使用方式

4. 持久化运行

English

Supported Link Types

Core Features

Installation

Configuration

Usage

CLI Mode (Desktop)

Bot Mode (Mobile)

1. Additional Configuration

2. Start Bot

3. How to Use

4. Keep Running

Supported LLM Providers

License

About

Topics

Resources

Uh oh!

Stars

Watchers

Forks

Releases

Packages 0

Contributors 2

Uh oh!

Languages

Packages