How to Build an AI Daily News Run with OpenClaw

Overview

An AI daily news run collects sources, summarizes stories, and delivers them to your team via Feishu. The pipeline has four stages: collect, summarize, format, and deliver.

Step 1: Install Required Skills

Install the skills needed for research, audio, and image generation:

openclaw skills install @clawhub/tavily
openclaw skills install @clawhub/listen-hub
openclaw skills install @clawhub/nano-banana-pro

• Tavily: Web search for news sources and trending topics
• ListenHub: Text-to-speech for audio briefings
• Nano Banana Pro: Image generation for daily cover images

Verify installation:

openclaw skills list

Step 2: Define News Sources

Create a config file that defines your news sources and topics:

# news-sources.yaml
sources:
  - name: tech
    feeds:
      - https://news.ycombinator.com/rss
      - https://feeds.arstechnica.com/arstechnica/index
  - name: ai
    feeds:
      - https://openai.com/blog/rss.xml
      - https://www.anthropic.com/news/rss

Adjust URLs to match your domain (AI, finance, legal, etc.). Keep the list small at first (3–5 sources) to avoid overload.

Step 3: Configure Output Format

The daily run can produce multiple outputs:

Output	Format	Skill / Method
Summary doc	Feishu Doc	OpenClaw Feishu app + doc API
Data table	Feishu Sheet	Sheet API with rows per story
Audio briefing	MP3	ListenHub TTS
Cover image	PNG	Nano Banana Pro

Define a prompt template that produces structured output:

Today's AI News Digest for {date}
---
Top 5 stories:
1. [Title] - [2-sentence summary] - [Source URL]
2. ...
---
Audio script (2 min): [Condensed version for TTS]
---
Cover image prompt: "Modern abstract illustration of AI and news, professional, minimal"

Step 4: Set Up Feishu Bot Delivery

1. Create a Feishu group for the daily digest
2. Add the OpenClaw Feishu bot to the group
3. Grant the bot permission to create docs and sheets in a shared drive
4. Configure the run to post the doc link and optionally the audio file to the group

Use the Feishu channel config:

channels:
  feishu:
    app_id: <your-app-id>
    app_secret: <your-secret>
    default_group_id: <digest-group-id>
    output_drive: <drive-id-for-docs>

Step 5: Schedule with Cron

Run the pipeline daily at 7:00 AM (adjust timezone as needed):

# Crontab entry (Linux/macOS)
0 7 * * * /usr/local/bin/openclaw run --config /path/to/daily-news.yaml --prompt "Generate daily news digest"

Or use a systemd timer, GitHub Actions (on: schedule: cron: '0 7 * * *'), or your cloud scheduler. Ensure OPENCLAW_API_KEY and other env vars are available to the cron environment.

Step 6: Test the Pipeline

Before relying on cron, run manually:

openclaw run --config daily-news.yaml --prompt "Generate today's AI news digest (test run)"

Check that:

1. Tavily returns results
2. Summaries are written to a Feishu doc
3. The sheet is populated
4. Audio is generated (if enabled)
5. The cover image is created

Common Issues

• Empty results: Verify Tavily API key and source URLs
• Feishu permission denied: Ensure the bot has doc:doc, doc:sheet, and im:message scopes
• Cron fails silently: Log output to a file: 0 7 * * * openclaw run ... >> /var/log/openclaw-news.log 2>&1
• Too many tokens: Limit to 5–10 stories per digest to stay within limits

For more skill ideas, see Best Skills for Beginners.