Problem
Marketplace sellers lose on price without knowing it: a competitor drops their price this morning, your product sinks to page two, and you only find out three days later when revenue has already tanked. On the other side, plenty panic and cut prices too — on margins that are already thin. Raw-material prices (palm oil, flour, resin, fabric) move every week, but the price list rarely gets touched. Pricing decisions come from gut feel and Seller Center screenshots, not from daily data.
Recognize this?
- Losing on price; first clue is a buyer's chat warning you
- Competitor cuts price today; you hear about it in three days
- Panic-matching price cuts when the margin is already thin
- Price lists edited from feeling and Seller Center screenshots
✕ Now
- Competitor price moves surface days later
- Spontaneous price reactions, no basis
- Price war only felt after revenue tanks
✦ With Mintaka
- Competitor prices watched daily; changes caught < 12 hours
- Daily per-SKU recommendations + reasons: margin floor, elasticity, market position
- Instant alarm when ≥2 competitors cut price within < 24 hours
What the agent does
The agent tracks chosen competitors' prices on marketplaces and public material/benchmark prices every day, then sends price-adjustment recommendations over WhatsApp — reasons included — plus an instant alarm when a price war is detected starting.
✦ Handled by the agent
- watches competitor and raw-material prices daily, compares them against your margins, and sends recommendations along with price-war alarms.
◆ Decided by you
- change the prices — via one-click approve — and set the margin floor.
Pipeline
your products mapped to competitor listings (match by title/variant/seller; LLM-assisted matching, ambiguous cases confirmed by a human). 10–50 SKU typical.
price, stock status, rating, estimated sold (sales score) from Shopee/Tokopedia/Lazada; material/benchmark prices from public sources (BPS data/commodity prices, industry sites). Respectful collection: rate-limit, jitter, cache fallback.
delta vs baseline: competitor down >3%, out of stock (chance to raise price/ad spend), flash sale, variant change. Price-war alarm: ≥2 competitors down >X% within <24 hours → instant notification, no waiting for the morning report.
LLM drafts the daily proposal per SKU: hold / drop to $Y / raise to $Z, with reasons (your margin floor, elasticity from history, competitor position) — executed by a human through Seller Center (or API if the client wants semi-automatic mode).
concise daily via WA (≤10 lines); weekly: your price position vs the market, margin held, recommendations followed vs ignored and the revenue result.
Integrations
- Shopee / Tokopedia / Lazada (public pages + seller API when available)
- Material price sources: public BPS/commodity data, industry association sites
- Output → WhatsApp owner (Fonnte/WATI/Meta Cloud API) + Mission Control; history to Google Sheet
Model stack
- GLM 5.3 Flash for product matching, price-position analysis, and drafting recommendations (prompt caching = product profile + your margin policy)
- Qwen3.8 Flash for reading product-page screenshots when the HTML structure changes / prices only appear as images
- Cost estimate: < US$15/month for 50 SKU × 10 competitors, 2x crawl/day
Success metrics (tracked weekly)
Sprint deliverables
SKU + competitor watchlist approved (mapping manually verified)
✓Daily crawler + delta-detection pipeline live
✓Daily WA recommendation format approved by the owner
✓Price-war alarm + threshold rules agreed with the client
✓Flow documentation + code handover (perpetual)
✓Ready to light up this star?
Free 30-minute discovery call: we'll show you exactly how many credits this takes for your business.
Book a Discovery Call →