AI News Briefing

Automated daily AI news research agent that searches the web, compiles a structured briefing, publishes it to Notion, and delivers a styled summary to Microsoft Teams and Slack -- powered by Claude Code CLI. Supports both macOS (launchd) and Windows (Task Scheduler). Includes an on-demand Custom Brief mode for deep multi-agent research on any topic. Fully automated pipeline with zero manual intervention required after setup.

[!NOTE]
Live AI News Notion page: https://hoangsonw.notion.site/9c34d052d9354beda82a3423e2d2f404?v=d43c53fe405c4896bfd95ad0cc22246f

Overview

AI News Briefing is a fully automated pipeline that runs every morning on your machine. It uses Claude Code in headless mode to act as a news research agent: searching the web across 9 AI-related topics, compiling the results into a two-tier briefing (TL;DR + full report), writing the finished page directly to a Notion database, and optionally posting a styled Adaptive Card summary to Microsoft Teams and Slack.

The entire process -- from triggering to publishing -- requires zero human intervention. You wake up, open Notion (or Teams or Slack), and your daily AI briefing is already there.

Custom Brief extends this with on-demand deep research: pick any topic, and 5 parallel research agents investigate it from different angles (breaking news, technical analysis, industry impact, trends, and policy). Results publish to Notion, Teams, and/or Slack -- or just print to your terminal.

Why it exists

Keeping up with AI news across models, tools, policy, funding, and open source is a full-time job. This project compresses that into an automated daily digest that covers 9 topic areas in a consistent format, delivered to your Notion workspace before you start your workday. When you need to go deeper on a specific topic, the custom brief gives you a comprehensive research report on demand.

Architecture

We leverage the scheduling capabilities of the OS (launchd on macOS, Task Scheduler on Windows) to trigger a shell script at a specific time each day. The script reads a prompt template, invokes the Claude Code CLI in headless mode, and the agentic prompt performs web searches, compiles results, and calls the Notion MCP tool to create a new page in the database.

flowchart TD
    subgraph Schedulers
        A1[macOS launchd]
        A2[Windows Task Scheduler]
    end

    A1 -->|8:00 AM daily| B1[briefing.sh]
    A2 -->|8:00 AM daily| B2[briefing.ps1]

    B1 -->|Reads| C[prompt.md]
    B2 -->|Reads| C

    B1 -->|Invokes| D[Claude Code CLI]
    B2 -->|Invokes| D

    D -->|Step 1: Search| E[WebSearch Tool]
    E -->|9 topics x multiple queries| F[Web Results]
    F -->|Step 2: Compile| G[Two-Tier Briefing]
    G -->|Step 3: Write| H[Notion MCP]
    H -->|Creates page| I[Notion Database]

    D -->|Step 4: Write card| J2[logs/YYYY-MM-DD-card.json]

    B1 -->|If Teams webhook set| T[notify-teams.sh]
    B2 -->|If Teams webhook set| T2[notify-teams.ps1]
    T -->|POST card.json| V[Teams Webhook]
    T2 -->|POST card.json| V
    V --> W[Teams Channel]

    B1 -->|If Slack webhook set| S1[notify-slack.sh]
    B2 -->|If Slack webhook set| S2[notify-slack.ps1]
    S1 -->|Convert + POST| SV[Slack Webhook]
    S2 -->|Convert + POST| SV
    SV --> SW[Slack Channel]

    B1 -->|Logs output| J[logs/YYYY-MM-DD.log]
    B2 -->|Logs output| J
    J -->|Auto-cleanup| K[Delete logs older than 30 days]

Data flow summary:

1. The platform scheduler fires the entry point script (briefing.sh on macOS, briefing.ps1 on Windows) at the configured time each day.
2. The script reads the prompt from prompt.md and passes it to the Claude Code CLI in print mode.
3. Claude Code executes the prompt as an agentic task -- performing web searches, compiling results, and calling the Notion MCP tool.
4. Notion receives the finished briefing as a new database page.
5. If the AI_BRIEFING_TEAMS_WEBHOOK environment variable is set, the entry point script calls notify-teams.sh / notify-teams.ps1, which validates and POSTs the pre-built logs/YYYY-MM-DD-card.json file (written by Claude in Step 4) to the configured Teams webhook.
6. If the AI_BRIEFING_SLACK_WEBHOOK environment variable is set, the entry point script calls notify-slack.sh / notify-slack.ps1, which converts the Teams card to Slack Block Kit format and POSTs it to the configured Slack webhook.
7. Logs are written to a date-stamped file and automatically pruned after 30 days.

Prerequisites

Installation

1. Clone the project

git clone https://github.com/hoangsonww/AI-News-Briefing
cd AI-News-Briefing

2. Platform-specific setup

macOS (launchd)

# Make the shell script executable
chmod +x ~/ai-news-briefing/briefing.sh

# Install the launchd plist
cp ~/ai-news-briefing/com.ainews.briefing.plist ~/Library/LaunchAgents/
launchctl load ~/Library/LaunchAgents/com.ainews.briefing.plist

# Verify the agent is registered
launchctl list | grep ainews

Optionally, set up the manual trigger command:

mkdir -p ~/.local/bin
cat > ~/.local/bin/ai-news << 'EOF'
#!/bin/bash
echo "Starting AI News Briefing..."
launchctl kickstart "gui/$(id -u)/com.ainews.briefing"
echo "Running. Check Notion or: tail -f ~/ai-news-briefing/logs/$(date +%Y-%m-%d).log"
EOF
chmod +x ~/.local/bin/ai-news

Make sure ~/.local/bin is in your PATH (add export PATH="$HOME/.local/bin:$PATH" to your ~/.zshrc if needed).

Windows (Task Scheduler)

Open PowerShell and run the installer script:

cd $env:USERPROFILE\ai-news-briefing
.\install-task.ps1

This registers a Task Scheduler task named AiNewsBriefing that runs daily at 8:00 AM under the current user account.

To customize the time:

.\install-task.ps1 -Hour 7 -Minute 30

Configuration

Change the schedule

macOS: Edit com.ainews.briefing.plist and modify the StartCalendarInterval section, then reload:

launchctl unload ~/Library/LaunchAgents/com.ainews.briefing.plist
launchctl load ~/Library/LaunchAgents/com.ainews.briefing.plist

Windows: Re-run the installer with new time parameters:

.\install-task.ps1 -Hour 9 -Minute 0

Change the model

Edit the entry point script for your platform:

• macOS: briefing.sh -- change the --model sonnet flag
• Windows: briefing.ps1 -- change the --model sonnet argument

Model trade-offs:

Change the budget cap

Edit the --max-budget-usd value in briefing.sh (macOS) or briefing.ps1 (Windows). The default is 2.00 (USD per run). This acts as a safety cap -- if the agent's token usage would exceed this amount, the run stops.

Change the topics

Edit prompt.md and modify the "Topics to Search" list. You can add, remove, or rename topics. If you change the number of topics, also update the "Topics": 9 value in the Notion properties section at the bottom of the prompt.

Usage

Automatic (scheduled)

Once installed, the briefing runs automatically every day at the scheduled time (default: 8:00 AM). No action needed.

Manual trigger

Using Make (recommended, cross-platform):

make run            # Run in foreground
make run-bg         # Run in background
make run-scheduled  # Trigger via OS scheduler

Platform-native:

# macOS
ai-news
# or: launchctl kickstart "gui/$(id -u)/com.ainews.briefing"

# Windows (PowerShell or cmd)
schtasks /run /tn AiNewsBriefing

Watch the progress

make tail           # Cross-platform: tail today's log

Or platform-native:

# macOS
tail -f ~/ai-news-briefing/logs/$(date +%Y-%m-%d).log

# Windows (PowerShell)
Get-Content "$env:USERPROFILE\ai-news-briefing\logs\$(Get-Date -Format 'yyyy-MM-dd').log" -Wait

A typical successful run takes 2-4 minutes and ends with a message like:

2026-03-09 14:08:08 Briefing complete. Check Notion for today's report.

Makefile Reference

The project includes a cross-platform Makefile that auto-detects your OS and routes commands to the correct platform tools. Requires GNU Make.

Utility Scripts Reference

The scripts/ directory contains 13 utility script pairs (.sh for macOS/Linux, .ps1 for Windows) for managing and troubleshooting the system.

Windows equivalents use the same names with .ps1 extension and PowerShell parameter syntax (e.g., .\scripts\health-check.ps1, .\scripts\topic-edit.ps1 -Action add -Name "AI Hardware" -Description "GPU news").

Notion Setup

Database schema

The prompt expects a Notion database with at least these properties:

You can add additional properties to the database (tags, priority, etc.), but the three above are what the agent writes to.

Data source ID

The prompt references a specific Notion data source ID:

856794cc-d871-4a95-be2d-2a1600920a19

To use your own database, replace this value in prompt.md (in the Step 3 section). To find your data source ID:

1. Open Claude Code and ensure the Notion MCP is connected.
2. Ask Claude: "List my Notion data sources" or use the notion-search MCP tool.
3. Copy the data_source_id for the database you want to use.
4. Replace the ID in prompt.md.

Page format

Each generated page contains:

• TL;DR -- 10-15 bullet points covering the biggest stories (roughly a 1-minute read)
• Divider
• Full Briefing -- 9 sections (one per topic), each with 3-8 detailed bullet points and source attribution
• Key Takeaways table -- a summary table of major trends and signals

Delivery Integrations (Teams + Slack)

The notification system supports both Microsoft Teams and Slack after a successful briefing run. Both channels are optional and can be enabled independently.

• Shared source artifact: Claude writes one card file, logs/YYYY-MM-DD-card.json, in Step 4.
• Teams path: notify-teams.sh/.ps1 validates JSON and POSTs the payload directly to Teams webhook URL(s).
• Slack path: notify-slack.sh/.ps1 converts the same card file to Slack Block Kit using scripts/teams-to-slack.py, then POSTs to Slack webhook URL(s).
• Fan-out support: Both channels support semicolon-separated webhook URLs.

flowchart LR
    A["Claude writes logs/YYYY-MM-DD-card.json"] --> B{"Channel"}
    B --> C["notify-teams.sh / notify-teams.ps1"]
    B --> D["notify-slack.sh / notify-slack.ps1"]
    C --> E["POST Adaptive Card JSON to Teams webhooks"]
    D --> F["teams-to-slack.py conversion"]
    F --> G["POST Block Kit JSON to Slack webhooks"]

Teams Setup

1. Create a Teams webhook (Power Automate workflow).
   Full guide: NOTIFY_TEAMS.md
2. Set AI_BRIEFING_TEAMS_WEBHOOK.
3. Run the briefing or call notify-teams directly.

Slack Setup

1. Create a Slack app incoming webhook at api.slack.com/apps.
2. Set AI_BRIEFING_SLACK_WEBHOOK.
3. Run the briefing or call notify-slack directly.

Set Webhook URL(s) Persistently

macOS / Linux

# Teams
export AI_BRIEFING_TEAMS_WEBHOOK="https://first-teams-webhook"

# Slack
export AI_BRIEFING_SLACK_WEBHOOK="https://hooks.slack.com/services/T.../B.../..."

Windows (PowerShell)

[Environment]::SetEnvironmentVariable("AI_BRIEFING_TEAMS_WEBHOOK", "https://first-teams-webhook", "User")
[Environment]::SetEnvironmentVariable("AI_BRIEFING_SLACK_WEBHOOK", "https://hooks.slack.com/services/T.../B.../...", "User")

Multiple Webhooks and --all / -All

Configure multiple URLs with semicolons:

export AI_BRIEFING_TEAMS_WEBHOOK="https://teams-webhook-1;https://teams-webhook-2"
export AI_BRIEFING_SLACK_WEBHOOK="https://slack-webhook-1;https://slack-webhook-2"

• notify-teams / notify-slack default: send to first URL only.
• --all (bash) / -All (PowerShell): send to all configured URLs.
• Current briefing.sh and briefing.ps1 call both notifiers with all URLs.

# First URL only
bash scripts/notify-teams.sh
bash scripts/notify-slack.sh

# All configured URLs
bash scripts/notify-teams.sh --all
bash scripts/notify-slack.sh --all

Test Delivery Directly

macOS / Linux

bash scripts/notify-teams.sh --all --card-file logs/2026-03-24-card.json
bash scripts/notify-slack.sh --all --card-file logs/2026-03-24-card.json

Windows (PowerShell)

.\scripts\notify-teams.ps1 -All -CardFile .\logs\2026-03-24-card.json
.\scripts\notify-slack.ps1 -All -CardFile .\logs\2026-03-24-card.json

For full setup and troubleshooting:

• NOTIFY_TEAMS.md
• NOTIFY_SLACK.md

Custom Brief: Deep Topic Research

In addition to the daily automated briefing, you can run a deep research briefing on any topic on demand. This spawns 5 parallel research agents that investigate the topic from different angles, then synthesizes findings into a comprehensive news briefing.

flowchart LR
    subgraph "Input"
        A["--topic 'AI in healthcare'"]
        B["--notion --teams --slack"]
    end
    A --> C[custom-brief.sh / .ps1]
    B --> C
    C -->|5 parallel agents| D[Deep Research]
    D --> E[Structured Briefing]
    E --> F[Terminal Output]
    E -->|optional| G[Notion]
    E -->|optional| H[Teams]
    E -->|optional| I[Slack]

Quick start

# Full research with all destinations
./custom-brief.sh --topic "AI in healthcare" --notion --teams --slack

# Terminal + Notion only
./custom-brief.sh -t "quantum computing" -n

# Interactive mode (prompts for topic and destinations)
./custom-brief.sh

# PowerShell
.\custom-brief.ps1 -Topic "AI regulation EU" -Notion -Teams

# Make
make custom-brief T="open source LLMs" NOTION=1 TEAMS=1

What it produces

• TL;DR -- 5-10 bullet points with key findings
• Thematic sections -- 3-6 sections organized by theme with linked citations and dates
• Key Trends & Outlook -- strategic implications table
• Sources -- numbered list of every URL cited

Every finding includes a clickable source link and publication date. Full details: CUSTOM_BRIEF.md

How the Prompt Works

The prompt (prompt.md) instructs Claude to execute four sequential steps within a single agentic session:

Step 1: Search for News

Claude uses the WebSearch tool to perform multiple searches per topic, targeting news from the past 24-48 hours. The search strategy includes date-qualified queries like "[topic] news today 2026-03-09" and company-specific queries.

Step 2: Compile the Briefing

Search results are synthesized into a two-tier format:

• Tier 1 (TL;DR): 10-15 one-sentence bullet points covering the top stories across all topics. Designed as a quick-scan summary.
• Tier 2 (Full Briefing): 9 sections with detailed coverage, source attribution, and a closing Key Takeaways table.

Step 3: Write to Notion

Claude checks whether a page for today already exists (captured during Step 0b). If one exists, it updates the page in place. If not, it creates a new page in the target database. This prevents duplicate pages when the briefing runs multiple times in a day.

Step 4: Generate Teams Card JSON

Claude writes a complete Adaptive Card JSON payload to logs/YYYY-MM-DD-card.json. This is the exact file that gets POSTed to the Teams webhook -- no parser or transformation sits between the AI output and Teams. The card includes a styled header, TL;DR bullets, topic sections, and an action button linking to the full Notion page.

Here is what a card looks like in Teams:

And in Slack:

Topic Coverage

Logs

Location

All logs are stored in the logs/ directory within the project:

Additionally, the logs/YYYY-MM-DD-card.json file contains the raw Adaptive Card JSON generated by Claude for Teams notifications.

Reading logs

macOS:

cat ~/ai-news-briefing/logs/$(date +%Y-%m-%d).log
tail -f ~/ai-news-briefing/logs/$(date +%Y-%m-%d).log

Windows:

Get-Content "$env:USERPROFILE\ai-news-briefing\logs\$(Get-Date -Format 'yyyy-MM-dd').log"
Get-Content "$env:USERPROFILE\ai-news-briefing\logs\$(Get-Date -Format 'yyyy-MM-dd').log" -Wait

Auto-cleanup

Logs older than 30 days are automatically deleted at the end of each run on both platforms. The macOS-specific launchd-stdout.log and launchd-stderr.log files are not date-stamped and may need periodic manual cleanup.

Tests

We have several test suites, with 247 non-blocking tests that verify syntax, structure, arg handling, template substitution, card JSON, notification error paths, and cross-platform portability. No external services are called.

# All bash tests (macOS / Linux / Git Bash)
bash tests/run-all.sh

# Individual suites
bash tests/test-custom-brief.sh
bash tests/test-daily-brief.sh
bash tests/test-notifications.sh
bash tests/test-portability.sh

# PowerShell (Windows)
powershell -ExecutionPolicy Bypass -File tests\test-all.ps1

Full documentation: TESTS.md

Troubleshooting

"Claude Code cannot be launched inside another Claude Code session"

This error occurs when the CLAUDECODE environment variable is set, which happens if you trigger the script from inside a Claude Code terminal session. Both briefing.sh and briefing.ps1 unset this variable automatically, but if you see this error:

• Make sure you are running the briefing from a regular terminal, not from within Claude Code.
• Verify the entry point script contains the unset CLAUDECODE / $env:CLAUDECODE = $null line.

Scheduled task does not run at the expected time

macOS (launchd):

• Mac was asleep: launchd will run the job when the Mac wakes up if the scheduled time was missed. If Power Nap is disabled or the lid was closed, the job may not fire until the next login.
• Powered off at scheduled time: The job is skipped entirely for that day.
• Agent not loaded: Verify with launchctl list | grep ainews. If missing, reload the plist.
• Path issues: The plist sets a custom PATH and HOME. If Claude is installed in a non-standard location, update the PATH in the plist.

Windows (Task Scheduler):

• Machine was off/asleep: StartWhenAvailable is enabled, so the task runs as soon as the machine wakes or the user logs in.
• Task not registered: Verify with schtasks /query /tn AiNewsBriefing. If missing, re-run install-task.ps1.
• Execution policy: The task action uses -ExecutionPolicy Bypass. If this is overridden by group policy, contact your IT admin or run briefing.ps1 manually.

Run succeeds but no Notion page appears

• Check that the Notion MCP is configured in Claude Code's MCP settings.
• Verify the data source ID in prompt.md matches a database your Notion integration has access to.
• Look at the log output -- Claude typically prints a Notion URL on success.

Budget exceeded

If the log shows the run stopped mid-way, the --max-budget-usd cap may have been reached. Increase the budget in the entry point script or switch to a cheaper model.

Multiple runs in the same day

Running the briefing multiple times in a day updates the existing Notion page rather than creating a duplicate. The agent checks for an existing page during Step 0b and updates it if found. Logs append to the same date-stamped file, so all runs for a given day are captured in one log.

Cost Estimate

With the default configuration (opus model, 9 topics):

Actual costs vary based on the volume of news, number of search queries, and briefing length. The --max-budget-usd 2.00 cap ensures no single run exceeds $2.00.

Project Structure

ai-news-briefing/
├── index.html                   # Landing page / project wiki
├── wiki/                        # Landing page assets
│   ├── style.css                # Styles
│   └── script.js                # Interactions
├── Makefile                     # Cross-platform task runner
├── scripts/                     # Utility scripts (sh + ps1 pairs)
│   ├── health-check.sh/.ps1     # Verify full setup
│   ├── log-summary.sh/.ps1      # Summarize recent runs
│   ├── log-search.sh/.ps1       # Search across all logs
│   ├── dry-run.sh/.ps1          # Test without writing to Notion
│   ├── test-notion.sh/.ps1      # Test Notion MCP connectivity
│   ├── cost-report.sh/.ps1      # Estimate API costs from logs
│   ├── export-logs.sh/.ps1      # Archive logs to tar.gz/zip
│   ├── backup-prompt.sh/.ps1    # Version and restore prompt.md
│   ├── topic-edit.sh/.ps1       # Add/remove/list topics
│   ├── update-schedule.sh/.ps1  # Change daily run time
│   ├── notify.sh/.ps1           # Send native OS notifications
│   ├── notify-teams.sh/.ps1     # Post briefing to Microsoft Teams
│   ├── notify-slack.sh/.ps1     # Post briefing to Slack
│   ├── teams-to-slack.py        # Convert Teams Adaptive Card JSON to Slack Block Kit
│   ├── build-teams-card.py      # Legacy card builder (not used in current flow)
│   └── uninstall.sh/.ps1        # Full cleanup and removal
├── briefing.sh                  # Daily briefing entry point (bash)
├── briefing.ps1                 # Daily briefing entry point (PowerShell)
├── prompt.md                    # Daily briefing agent prompt
├── custom-brief.sh              # Custom topic briefing entry point (bash)
├── custom-brief.ps1             # Custom topic briefing entry point (PowerShell)
├── prompt-custom-brief.md       # Custom briefing deep research prompt
├── commands/
│   ├── ai-news-briefing.md      # Claude Code skill: daily briefing
│   └── custom-brief.md          # Claude Code skill: custom topic briefing
├── com.ainews.briefing.plist    # macOS launchd schedule definition
├── install-task.ps1             # Windows Task Scheduler installer
├── tests/                       # 247 non-blocking tests
│   ├── run-all.sh               # Bash test runner
│   ├── test-custom-brief.sh     # Custom brief tests
│   ├── test-daily-brief.sh      # Daily briefing tests
│   ├── test-notifications.sh    # Notification pipeline tests
│   ├── test-portability.sh      # Cross-platform portability tests
│   └── test-all.ps1             # PowerShell test suite
├── logs/                        # Run logs (git-ignored)
├── backups/                     # Prompt backups (git-ignored)
├── .gitignore
├── ARCHITECTURE.md              # Detailed architecture documentation
├── E2E_FLOW.md                  # End-to-end pipeline walkthrough
├── CUSTOM_BRIEF.md              # Custom topic briefing documentation
├── TESTS.md                     # Test suite documentation
├── LOGS.md                      # Log tailing and management guide
├── SETUP.md                     # Full setup guide
├── NOTIFY_TEAMS.md              # Teams integration setup guide
├── NOTIFY_SLACK.md              # Slack integration setup guide
└── README.md                    # This file

Author

Created by Son Nguyen -- AI researcher and developer focused on building tools that empower people to harness the power of AI in their daily lives.

Thanks for checking out the project! If you have any questions, suggestions, or want to contribute, feel free to open an issue or submit a pull request.