claudeers.

🔓 unclaimed — this page was auto-generated from GitHub. Are you the creator?

Claim this page →
// Automation & Workflows

ai-dispatch

Your daily AI intelligence dispatch to Email 📧· Robotics, Agents & LLMs analyzed by Claude Opus · 每日倚源聚合 + 深床分析GitHub Actions 䞀键郚眲无需服务噚

// Automation & Workflows[ cli ][ api ][ web ][ claude ]#claude#ai#ai-agents#anthropic#automation#email-automation#github-actions#llm◷ MIT$open-sourceupdated 15 days ago
Actively maintained
100/100
last commit 2 days ago
last release about 1 month ago
releases 1
open issues 1
// install
git clone https://github.com/Yifannnnnnnnw/ai-dispatch

📡 AI Dispatch

Your daily AI intelligence briefing, delivered to your inbox.

Automatically aggregates the latest in AI, Robotics, and Agents every morning — analyzed by an LLM of your choice, delivered to your inbox. Runs entirely on GitHub Actions. No server. No subscription.

Workflow


What You Get

Every email contains five structured sections:

SectionContent
📌 Top Stories10–15 curated items, each with significance analysis and cross-story connections
📈 Trend AnalysisCross-article patterns with evidence and forward predictions
🔬 Papers Worth ReadingSelected arXiv papers with core contributions and reading focus
📖 Blog PickOne deep-read recommendation (never repeats, auto-deduped)
💡 Today's SignalThe one judgment that matters most today, in one sentence

View a sample email →


Quick Start

No terminal required — everything runs in your browser.

Prerequisites

  • GitHub account (free)
  • Gmail account

Step 1 — Fork this repo

Click Fork in the top right → create it under your own account.


Step 2 — Run the Setup workflow

Go to Actions → ⚙ Setup → Run workflow and fill in the form:

FieldWhat to enter
LLM providergemini (free) or anthropic (paid)
Gmail addressyour Gmail address
Recipient emailleave blank to use the Gmail address above
Send time (UTC)hour 0–23 — Beijing 08:00 → 0, London BST 07:00 → 6, New York 07:00 → 11
Output languageEnglish or äž­æ–‡

The workflow updates config.yml and prints a checklist of the secrets you need to add next.


Step 3 — Add secrets

Go to Settings → Secrets and variables → Actions → New repository secret

Add these 4 secrets (the Setup workflow tells you exactly what to put in each):

SecretValue
GEMINI_API_KEYFree key from aistudio.google.com/apikey — or ANTHROPIC_API_KEY if you chose Anthropic
GMAIL_USERYour Gmail address
GMAIL_APP_PASSWORD16-character app password — how to get one ↓
RECIPIENT_EMAILDestination inbox

Gmail App Password

Gmail requires an app-specific password, not your account password.

  1. Go to myaccount.google.com/security
  2. Confirm 2-Step Verification is enabled
  3. Search App Passwords → open it
  4. Select Mail + Mac → click Generate
  5. Copy the 16-character password (shown only once, no spaces)

Step 4 — Verify

Go to Actions → ✅ Check Setup → Run workflow

── GitHub Secrets ──────────────────────────────────
  ✅  GEMINI_API_KEY     (set)
  ✅  GMAIL_USER         (set)
  ✅  GMAIL_APP_PASSWORD (set)
  ✅  RECIPIENT_EMAIL    (set)

── config.yml ──────────────────────────────────────
  ✅  config.yml found
  ✅  topics configured  (3 topics)
  ✅  news_feeds configured  (9 sources)
  ✅  blog_feeds configured  (8 blogs)

── LLM API ──────────────────────────────────────────
  ✅  API connection successful

── Gmail SMTP ───────────────────────────────────────
  ✅  Gmail login successful ([email protected])

── Test email ───────────────────────────────────────
  ✅  Test email sent (check your inbox)

══════════════════════════════════════════════════════
  🎉  All checks passed! Your daily digest starts tomorrow.
══════════════════════════════════════════════════════

Once all green, AI Dispatch runs automatically every day. The default send time targets 07:00 BST / 07:00 GMT — change it via send_hour_utc in config.yml.


Prefer the command line?

Set up locally with the interactive wizard (requires Git, Python 3.10+, and GitHub CLI).

Step 0 — Install Git and GitHub CLI

Install Git

# macOS — comes pre-installed; if missing:
xcode-select --install
# Windows
winget install Git.Git
# Linux (Debian / Ubuntu)
sudo apt install git

Windows: After winget installs Git, close and reopen your terminal before continuing.

Install GitHub CLI

# macOS
brew install gh
# Windows — open a new terminal after this completes
winget install GitHub.cli
# Linux (Debian / Ubuntu)
sudo apt install gh

Windows: Same as above — reopen your terminal after installation so gh is on your PATH.

Log in to GitHub

gh auth login

Follow the prompts — select GitHub.com → HTTPS → Login with a web browser.

Step 1 — Fork, clone, and launch

# macOS / Linux
gh repo fork Yifannnnnnnnw/ai-dispatch --clone
cd ai-dispatch        # use the folder name printed by gh above
python3 setup.py
# Windows
gh repo fork Yifannnnnnnnw/ai-dispatch --clone
cd ai-dispatch        # use the folder name printed by gh above
python setup.py

gh prints the local path after cloning, e.g. Cloned fork's Git repository to ai-dispatch.

The wizard asks a few questions and handles everything else — secrets, config, and push.

Step 2 — Verify

Go to Actions → ✅ Check Setup → Run workflow and confirm all checks pass.


Cost

GitHub Actions is always free. The only potential cost is the LLM API call, depending on which provider you use.

Default (Gemini — free):

ModelCostNotes
gemini-2.0-flashFreeDefault — 1,500 requests/day limit, more than enough for a daily digest
gemini-2.5-proFree within quotaHigher quality, with per-minute rate limits

Optional (Anthropic — paid):

ModelPer dayPer monthNotes
claude-sonnet-4-6~$0.05~$1.50Great quality
claude-opus-4-7~$0.67~$20Highest quality

To switch providers, update provider and digest.model in config.yml.


File Structure

ai-dispatch/
├── config.yml              ← Your personalization (the only file to edit)
├── setup.py                ← Interactive setup wizard
├── fetch_news.py           ← Main pipeline
├── check_setup.py          ← Setup verification script
├── requirements.txt
├── sent_history.json       ← Auto-maintained dedup log (do not edit manually)
└── .github/workflows/
    ├── daily_news.yml      ← Daily cron job
    ├── setup.yml           ← First-time setup wizard (browser-based)
    └── check_setup.yml     ← One-click setup check

FAQ

Q: Test email arrived but no daily digest? Check Actions → AI Dispatch for errors. GitHub Actions cron can occasionally delay 15–30 minutes.

Q: Gmail login fails (SMTPAuthenticationError)? Make sure you're using the 16-character app password, not your Gmail account password.

Q: How do I change the output language? Edit output_language in config.yml. Default is English — change it to äž­æ–‡ for Chinese output. The setup wizard also lets you choose during initial setup.

Q: How do I add my own RSS sources? Add a line under news_feeds or blog_feeds in config.yml: Source Name: https://rss-url.

Q: Blog picks keep repeating? sent_history.json tracks all previously sent URLs. To reset, clear the urls array in that file.



📡 AI Dispatch䞭文

每倩早䞊AI 驱劚的深床简报自劚聚合分析发到䜠的邮箱。

党皋运行圚 GitHub Actions 䞊䞍需芁服务噚䞍需芁订阅莹Fork 即甚。


效果预览

每封邮件包含五䞪固定板块

板块内容
📌 重点新闻10–15 条粟选每条附意义分析和关联刀断
📈 趋势分析跚文章園纳的行䞚/技术趋势及预刀
🔬 倌埗深挖粟选 arXiv 论文诎明栞心莡献和阅读重点
📖 今日掚荐博客1 篇深床富读自劚去重䞍重倍
💡 今日信号䞀句话最关键刀断

查看瀺䟋邮件 →


快速匀始

前眮条件

  • GitHub 莊号免莹
  • Gmail 莊号

党皋圚浏览噚完成无需安装任䜕蜯件。


第䞀步Fork 仓库

点击右䞊角 Fork → 创建到䜠自己的莊号䞋。


第二步运行 Setup workflow

进入仓库 → Actions → ⚙ Setup → Run workflow填写衚单

字段填写内容
LLM providergemini免莹或 anthropic付莹
Gmail 地址䜠的 Gmail 地址
收件邮箱留空则默讀䞎 Gmail 地址盞同
发送时闎UTC小时 0–23 — 北京 08:00 → 0䌊敊 BST 07:00 → 6纜纊 07:00 → 11
蟓出语蚀English 或 äž­æ–‡

workflow 运行完成后䌚自劚曎新 config.yml并圚日志䞭打印需芁添加的 Secrets 枅单。


第䞉步添加 Secrets

进入仓库 → Settings → Secrets and variables → Actions → New repository secret

按 Setup workflow 日志䞭的提瀺添加以䞋 4 䞪 Secrets

Secret 名称填写内容
GEMINI_API_KEY圚 aistudio.google.com/apikey 免莹申请若选 Anthropic 则改䞺 ANTHROPIC_API_KEY
GMAIL_USER䜠的 Gmail 地址
GMAIL_APP_PASSWORD16 䜍应甚密码 — 获取方法 ↓
RECIPIENT_EMAIL收件邮箱

Gmail 应甚密码

Gmail 䞍允讞盎接甚莊号密码需芁生成䞓甚的「应甚密码」。

  1. 打匀 myaccount.google.com/security
  2. 确讀䞀步验证已匀启未匀启则先匀启
  3. 搜玢框蟓入 App Passwords → 进入
  4. 选择「邮件」+「Mac」→ 点击生成
  5. 倍制星瀺的 16 䜍密码只星瀺䞀次去掉空栌填入

第四步验证配眮

进入仓库 → Actions → ✅ Check Setup → Run workflow

── GitHub Secrets ──────────────────────────────────
  ✅  GEMINI_API_KEY     (已讟眮)
  ✅  GMAIL_USER         (已讟眮)
  ✅  GMAIL_APP_PASSWORD (已讟眮)
  ✅  RECIPIENT_EMAIL    (已讟眮)

── config.yml ──────────────────────────────────────
  ✅  config.yml 存圚
  ✅  topics 已配眮      (3 䞪䞻题)
  ✅  news_feeds 已配眮  (9 䞪来源)
  ✅  blog_feeds 已配眮  (8 䞪博客)

── LLM API ──────────────────────────────────────────
  ✅  API 连接成功

── Gmail SMTP ───────────────────────────────────────
  ✅  Gmail 登圕成功 ([email protected])

── 测试邮件 ─────────────────────────────────────────
  ✅  测试邮件已发送 (请检查收件箱)

══════════════════════════════════════════════════════
  🎉  所有检查通过查收测试邮件后即可等埅每日简报。
══════════════════════════════════════════════════════

党郚绿色后每倩自劚运行默讀目标到蟟时闎䞺 BST 07:00 / GMT 07:00。


偏奜呜什行

䜿甚本地亀互向富配眮需芁 Git、Python 3.10+ 和 GitHub CLI。

第零步安装 Git 和 GitHub CLI

安装 Git

# macOS — 䞀般已预装劂果没有
xcode-select --install
# Windows
winget install Git.Git
# LinuxDebian / Ubuntu
sudo apt install git

Windows 泚意 winget 安装完后关闭终端重新打匀再继续。

安装 GitHub CLI

# macOS
brew install gh
# Windows — 安装完成后需芁重新打匀终端
winget install GitHub.cli
# LinuxDebian / Ubuntu
sudo apt install gh

Windows 泚意 同䞊安装后重新打匀终端gh 才胜被识别。

登圕 GitHub

gh auth login

按提瀺选择 GitHub.com → HTTPS → Login with a web browser。

第䞀步Fork、clone 并启劚向富

# macOS / Linux
gh repo fork Yifannnnnnnnw/ai-dispatch --clone
cd ai-dispatch        # 甹 gh 蟓出的文件倹名通垞就是 ai-dispatch
python3 setup.py
# Windows
gh repo fork Yifannnnnnnnw/ai-dispatch --clone
cd ai-dispatch        # 甹 gh 蟓出的文件倹名通垞就是 ai-dispatch
python setup.py

向富䌚自劚写入所有 Secrets、曎新 config.yml 并掚送。

第二步验证配眮

进入仓库 → Actions → ✅ Check Setup → Run workflow确讀所有检查通过。


莹甚参考

GitHub Actions 完党免莹。唯䞀的可胜成本是 LLM API 调甚取决于䜠选择的 provider。

Gemini默讀免莹

暡型莹甚诎明
gemini-2.0-flash免莹默讀每倩 1500 次请求限额日报绰绰有䜙
gemini-2.5-pro免莹额床内免莹莚量曎高有每分钟请求限制

Anthropic付莹可选

暡型每倩纊每月纊诎明
claude-sonnet-4-6Â¥0.36Â¥11莚量埈奜
claude-opus-4-7Â¥4.80Â¥144最高莚量

切换 provider 只需改 config.yml 侭的 provider 和 digest.model 字段。


文件诎明

ai-dispatch/
├── config.yml              ← 䜠的䞪性化配眮唯䞀需芁猖蟑的文件
├── setup.py                ← 亀互匏配眮向富
├── fetch_news.py           ← 䞻皋序
├── check_setup.py          ← 配眮验证脚本
├── requirements.txt
├── sent_history.json       ← 已掚送博客记圕自劚绎技请勿手劚猖蟑
└── .github/workflows/
    ├── daily_news.yml      ← 每日定时任务
    └── check_setup.yml     ← 䞀键验证配眮

垞见问题

Q: 测试邮件收到了䜆每日邮件没来 检查 Actions → AI Dispatch 里有没有报错。GitHub Actions 的 cron 有时䌚延迟 15–30 分钟。

Q: Gmail 登圕倱莥 (SMTPAuthenticationError) 确讀甚的是「应甚䞓甚密码」16䜍而䞍是 Gmail 莊号密码本身。

Q: 劂䜕切换蟓出语蚀 修改 config.yml 侭的 output_language 字段。默讀䞺 English改䞺 äž­æ–‡ 即蟓出䞭文。配眮向富䞭也可以圚初始讟眮时选择。

Q: 劂䜕添加自己的 RSS 源 圚 config.yml 的 news_feeds 或 blog_feeds 䞋新增䞀行名称: RSS铟接。

Q: 掚荐博客䞀盎重倍 sent_history.json 记圕已掚送内容劂需重眮枅空该文件的 urls 数组即可。

// compatibility

Platformscli, api, web
Operating systems—
AI compatibilityclaude
LicenseMIT
Pricingopen-source
LanguagePython

// faq

What is ai-dispatch?

Your daily AI intelligence dispatch to Email 📧· Robotics, Agents & LLMs analyzed by Claude Opus · 每日倚源聚合 + 深床分析GitHub Actions 䞀键郚眲无需服务噚. It is open-source on GitHub.

Is ai-dispatch free to use?

ai-dispatch is open-source under the MIT license, so it is free to use.

What category does ai-dispatch belong to?

ai-dispatch is listed under automation in the Claudeers registry of Claude-compatible tools.

0 views
★ 20 stars
unclaimed
updated 15 days ago

// embed badge

ai-dispatch on Claudeers
[![Claudeers](https://claudeers.com/api/badge/ai-dispatch.svg)](https://claudeers.com/ai-dispatch)

// retro hit counter

ai-dispatch hit counter
[![Hits](https://claudeers.com/api/counter/ai-dispatch.svg)](https://claudeers.com/ai-dispatch)

// reviews

// guestbook

0/500

// related in Automation & Workflows

🔓

The agent that grows with you

// automationNousResearch/⟹Python⟩★ 211,605◷ MIT[ claude ]
🔓

The API to search, scrape, and interact with the web at scale. 🔥

// automationfirecrawl/⟹TypeScript⟩★ 143,720◷ AGPL-3.0[ claude ]
🔓

🌐 Make websites accessible for AI agents. Automate tasks online with ease.

// automationbrowser-use/⟹Python⟩★ 103,709◷ MIT[ claude ]
🔓

An open-source long-horizon SuperAgent harness that researches, codes, and creates. With the help of sandboxes, memories, tools, skill, subagents and message


// automationbytedance/⟹Python⟩★ 76,016◷ MIT[ claude ]
→ see how ai-dispatch connects across the ecosystem