Responsible use & legal notes
What you can monitor
Price intelligence drives two critical decisions: whether you're undercut, and where you stand in the market. OpenClaw automates the monitoring, so you can react before a competitor's move costs you sales.
Price change detection
Alert when a competitor changes the price of a tracked product — immediately flag whether they went up or down.
Undercut alerts
Immediate flag if a competitor drops below your price — perfect for price-sensitive categories where even 5% matters.
Price gap analysis
Weekly report showing where you're priced high/low vs competitors — grouped by category with flags for overpriced and underpriced products.
Price history tracking
Log prices over time to spot seasonal patterns, promotion cycles, and competitor behavior — historical data beats guessing.
New product detection
Alert when a competitor adds a product in your category — catch new competition before they take market share.
How price scraping works
OpenClaw uses the http-crawler tool to fetch competitor product pages on a schedule. It extracts the price using either a CSS selector (for pages with classic HTML) or JSON-LD ProductPrice schema (for modern structured data), stores the previous price in a state file, and compares on each run. If the price has changed, an alert is triggered. If the extraction fails (page structure changed, selector broke), the agent flags it as a PARSE_ERROR instead of silently returning stale data.
The basic flow is:
- Fetch: HTTP GET request to the competitor's product page (with request delay and user-agent to be respectful)
- Extract: Parse the HTML or JSON response using your configured CSS selector or schema fallback
- Store: Save the current price to state file alongside a timestamp
- Compare: Check if price differs from last check — if yes, flag as PRICE_CHANGE
- Alert: If change is significant (or meets your alert thresholds), include in the report
Setting up price tracking config
Create a price-watcher agent in your AGENTS.md file. The agent needs a list of competitor products to track, where to extract the price from on each page, and optional thresholds for alerting.
# AGENTS.md — price-watcher agent
agents:
price-watcher:
description: "Monitor competitor prices and alert on changes and undercuts"
tools:
- http-crawler
config:
request_delay_seconds: 3 # be respectful: 3 sec between requests
user_agent: "Mozilla/5.0 (compatible; your-store-price-monitor)"
competitor_products:
- url: "https://competitor-a.com/products/classic-tee-black"
product_name: "Classic Tee — Black"
your_sku: "SKU-0001"
your_price_source: "shopify" # get your price from your Shopify store
css_selector: ".product-price" # fallback CSS selector
jsonld_fallback: true # try JSON-LD ProductPrice schema first
- url: "https://competitor-b.com/shop/hoodie-navy"
product_name: "Hoodie — Navy"
your_sku: "SKU-0042"
your_price_source: "shopify"
css_selector: "span[data-price]"
jsonld_fallback: true
- url: "https://competitor-c.com/products/cap-logo"
product_name: "Cap — Logo"
your_sku: "SKU-0089"
your_price_source: "shopify"
css_selector: ".price-main"
jsonld_fallback: true
state_file: "price_history.json"
alert_on:
- price_decrease # competitor lowers price
- price_increase # competitor raises price
undercut_threshold: 0 # alert if competitor price ≤ your price
output:
format: markdown_table
columns:
- product
- your_price
- competitor_price
- difference
- difference_pct
- change_from_last_check
- alert_level
sort_by: difference_pct # worst (most undercut) firstKey fields:
request_delay_seconds:How long to wait between requests (3 sec is polite)css_selector:CSS selector to extract price (e.g..product-price,span[data-price])jsonld_fallback:If true, tries JSON-LD ProductPrice schema before CSS selectoryour_price_source:Where OpenClaw gets your price — typicallyshopifyorwoocommerceundercut_threshold:Percentage or absolute value — 0 means alert if competitor ≤ your price
Detecting price changes
Every time the price-watcher agent runs, it compares the current price against the stored price in price_history.json. If there's a difference, it's flagged. You can set thresholds to ignore tiny fluctuations (e.g. only alert if difference > $0.50).
# Example price_history.json (state file)
{
"SKU-0001": {
"product_name": "Classic Tee — Black",
"last_checked": "2026-03-25T06:00:00Z",
"last_price": 24.99,
"current_price": 24.99,
"change_from_previous": 0,
"price_history": [
{ "date": "2026-03-18T06:00:00Z", "price": 24.99 },
{ "date": "2026-03-11T06:00:00Z", "price": 26.99 },
{ "date": "2026-03-04T06:00:00Z", "price": 26.99 }
]
},
"SKU-0042": {
"product_name": "Hoodie — Navy",
"last_checked": "2026-03-25T06:00:00Z",
"last_price": 49.99,
"current_price": 44.99,
"change_from_previous": -5.00,
"alert_level": "UNDERCUT",
"price_history": [
{ "date": "2026-03-25T06:00:00Z", "price": 44.99 },
{ "date": "2026-03-18T06:00:00Z", "price": 49.99 }
]
}
}Price gap analysis
Beyond individual price changes, it's useful to see your overall pricing position: are you generally overpriced, underpriced, or competitive? The gap analysis agent aggregates all tracked products and calculates the price gap as a percentage:
# Gap calculation formula
# gap_pct = (your_price - avg_competitor_price) / avg_competitor_price × 100
#
# Example:
# your_price = $25.00
# avg_competitor_price = $22.50
# gap_pct = (25 - 22.5) / 22.5 × 100 = +11% (you are 11% overpriced)
#
# Interpretation:
# gap_pct > +15% → OVERPRICED — risk of losing price-sensitive buyers
# -10% to +15% → COMPETITIVE — pricing is in line with market
# gap_pct < -10% → UNDERPRICED — potential margin opportunityResults are grouped by product category, so you can see pricing position by segment:
# AGENTS.md — price-gap-analyzer agent
agents:
price-gap-analyzer:
description: "Analyze price gaps by category and flag overpriced/underpriced products"
tools:
- http-crawler
config:
state_file: "price_history.json"
grouping: by_category # group results by product category tag
gap_thresholds:
overpriced: 15 # flag if gap > 15%
underpriced: -10 # flag if gap < -10%
min_competitors_to_analyze: 2 # need at least 2 competitors for meaningful gap
output:
format: markdown_table
columns:
- category
- product
- your_price
- avg_competitor_price
- gap_pct
- gap_flag
sort_by: gap_pct # most overpriced firstHEARTBEAT.md templates
Weekly price check
# HEARTBEAT.md — weekly price monitoring
schedule:
- name: "Weekly Price Check"
cron: "0 6 * * 1" # Mondays at 6 AM
agents:
- price-watcher
- price-gap-analyzer
output:
file: "weekly_price_report.md"
notify: true
prompt: |
Run the price-watcher and price-gap-analyzer agents. Format the output as:
1. PRICE CHANGES: list products where price changed since last check
2. UNDERCUTS: highlight any competitors now cheaper than us
3. OVERPRICED PRODUCTS: products where gap > 15%
4. UNDERPRICED PRODUCTS: products where gap < -10%
5. COMPETITIVE: products within -10% to +15% of average competitor
Keep it scannable. Use emoji for status: 🟢 competitive, 🟡 caution, 🔴 action needed.Immediate undercut alert (event-based)
# HEARTBEAT.md — real-time undercut alert
schedule:
- name: "Undercut Alert"
trigger: on_price_change # trigger when price-watcher detects a change
condition: "alert_level == 'UNDERCUT'" # only if competitor is now cheaper
agents:
- price-watcher
output:
file: "undercut_alert_[product].md"
notify: true
prompt: |
A competitor has dropped below our price. Output:
- Product name and SKU
- Their new price and our current price
- Percentage gap
- Action recommended (adjust price down, margin accept it, monitor only)What the weekly report looks like
Frequently asked questions
Is scraping competitor prices legal?
It exists in a legal grey area depending on your jurisdiction and the target site's terms of service. The data is publicly visible — any customer can see it — but automated collection may violate a site's ToS. Always check the target site's terms before monitoring. This guide covers monitoring publicly listed prices only, not accessing account areas, paywalled content, or internal pricing data.
What if the competitor changes their page structure?
If the CSS selector you configured breaks (because the competitor updated their HTML), the price extraction returns null. The agent flags this as a PARSE_ERROR rather than silently failing or returning a stale price. You then update the CSS selector in your config to match the new page structure.
How often should I check competitor prices?
Weekly is a good baseline for most products. Highly competitive categories may warrant twice-weekly checks. Avoid daily scraping — it strains the target server and increases the risk of your server IP getting blocked. Use sensible request delays (2–3 seconds between requests) and respect robots.txt rules.
Can OpenClaw track prices on Amazon?
Amazon actively blocks scrapers. If you're an Amazon Associate, use Amazon's Product Advertising API instead. For non-associates, consider commercial data providers like Keepa or DataForSEO's Amazon API endpoints. Standard web scraping of Amazon will result in IP blocks.