One-click end-to-end workflow for monitoring AI product launches. Runs four stages in sequence: RSS monitoring → product search → screenshot capture → trend analysis.
# Full pipeline (one click)
python scripts/run_pipeline.py
# Skip screenshot stage (no Playwright needed)
python scripts/run_pipeline.py --skip-screenshot
# Run a single stage
python scripts/run_pipeline.py --stage rss
python scripts/run_pipeline.py --stage search
python scripts/run_pipeline.py --stage screenshot
python scripts/run_pipeline.py --stage analysis
| # | Stage | What it does | Output |
|---|---|---|---|
| --- | ------- | ------------- | -------- |
| 1 | RSS Monitor | Fetches configured RSS feeds, detects new AI launch posts | data/raw_launches.json |
| 2 | Product Search | Enriches each launch with DuckDuckGo search results | data/enriched_launches.json |
| 3 | Screenshot | Captures full-page screenshots of product pages (optional) | screenshots/*.png, data/screenshot_results.json |
| 4 | Trend Analysis | Identifies keywords, top organizations, source distribution | analysis/trends.json, analysis/launch_analysis_report.md |
Edit config/rss_feeds.yaml to add or remove RSS feeds:
feeds:
- name: OpenAI Blog
url: https://openai.com/blog/rss.xml
category: llm
| Variable | Default | Description |
|---|---|---|
| ---------- | --------- | ------------- |
PIPELINE_BASE_DIR | skill root | Base directory for all outputs |
PIPELINE_DATA_DIR | {BASE}/data | JSON data output directory |
PIPELINE_SCREENSHOT_DIR | {BASE}/screenshots | Screenshot output directory |
PIPELINE_ANALYSIS_DIR | {BASE}/analysis | Analysis report directory |
PIPELINE_CONFIG | config/rss_feeds.yaml | RSS feed config path |
pip install pyyamlpip install playwright && playwright install chromiumAfter a full run:
data/
raw_launches.json # Stage 1 output
seen_ids.json # Dedup state
enriched_launches.json # Stage 2 output
screenshot_results.json # Stage 3 output
screenshots/
*.png # Captured screenshots
analysis/
trends.json # Structured trend data
launch_analysis_report.md # Human-readable report
Pair with OpenClaw cron for automated daily runs:
# Add via cron tool
schedule: { kind: "cron", expr: "0 8 * * *" }
payload: { kind: "agentTurn", message: "Run the AI launch pipeline: python ~/.openclaw/skills/ai-launch-pipeline/scripts/run_pipeline.py --skip-screenshot, then summarize the analysis report." }
共 2 个版本