HubesHub Automation Plan

A plan to automate pricing, inventory, and growth for HubesHub — while staying 100% within eBay's and Shopify's rules.

What We're Building

HubesHub has strong foundations: 99.9% eBay feedback, 3.1K items sold, $198K combined revenue across two channels, and deep knowledge of OEM auto parts for Ford, Dodge, Jeep, and Chevy. The goal is to layer automation on top of that expertise to:

Price smarter — Use eBay's own market data (via their official APIs) to set competitive, profitable prices on both eBay and Shopify
Manage inventory efficiently — Agents that flag stale listings, suggest repricing, and draft new listings from photos for both platforms simultaneously
Drive direct traffic — SEO and AI Engine Optimization (AEO) on Shopify to reduce eBay fee dependency over time
Source profitably — Data-driven shopping lists when visiting car yards, showing which parts are worth pulling

Everything we build will use official APIs from both eBay and Shopify, staying within their Terms of Service. No scraping, no bots that take autonomous actions, no shortcuts that risk either account.

Platform Compliance Rules

These rules govern every decision we make across both platforms. They come from eBay's API License Agreement (June 2025), eBay's User Agreement (February 2026), Shopify's API Terms (February 2026), and Shopify's Acceptable Use Policy.

Shared Rules (Both Platforms)

✓ Safe — what we will do

Use official APIs only to manage our own inventory, listings, orders, and analytics on both platforms. No scraping, no screen automation, no unofficial endpoints.

✓ Safe — what we will do

Human-in-the-loop for all actions. Every listing, price change, and bulk operation is a draft that Chris reviews and approves before going live. No autonomous actions on either platform.

✓ Safe — what we will do

Keep customer data private. We won't export names, emails, or addresses to external systems. Our automation focuses on products and inventory, not personal data.

⚠ Caution — allowed with restrictions

Respect rate limits and API versions. Both platforms throttle API requests. We'll pace our automation, wait when asked, run bulk jobs off-peak, and proactively update when new API versions are released (Shopify retires versions after ~12 months).

⚠ Caution — allowed with restrictions

Store platform data temporarily. Caching data for processing is acceptable on both platforms. We will not build permanent archives of competitor data or aggregate it beyond managing our own business.

✕ Prohibited — we will never do this

No scraping either platform — not our stores, not competitors, not marketplaces. Both eBay and Shopify explicitly prohibit "any robot, spider, scraper, or other automated means" to access their sites outside official APIs.

✕ Prohibited — we will never do this

No automatic cross-platform repricing without review. We will not automatically mirror eBay prices to Shopify or vice versa. Cross-platform auto-repricing is a compliance risk on both platforms. Chris reviews every price change.

✕ Prohibited — we will never do this

No feeding store data into AI models for training. AI helps us build the tools — it doesn't ingest product, customer, or order data for model training on either platform.

eBay-Specific Rules

✓ Safe — eBay specific

Use Product Research (Terapeak) in Seller Hub to look up sold prices, sell-through rates, and market trends. This is eBay's own tool, free for all sellers.

⚠ Caution — eBay specific

Automated repricing of our own eBay listings based on eBay market data is allowed, but we must NOT auto-reprice in response to prices on other platforms (Amazon, Car-Part.com, etc.). eBay calls this "seller arbitrage" and prohibits it.

⚠ Caution — eBay specific

AI and eBay "Restricted API" data. Using AI to write code is fine. Feeding eBay's Restricted API data (market trends, pricing strategies, sales volumes) into a third-party AI model for training requires eBay's written consent. Our pricing algorithms will be traditional code.

✕ Prohibited — eBay specific

No AI buy-for-me agents. eBay's Feb 2026 User Agreement bans "buy-for-me agents, LLM-driven bots, or any end-to-end flow that attempts to place orders without human review."

✕ Prohibited — eBay specific

No seller arbitrage automation. No tools that auto-reprice eBay listings based on other marketplaces, auto-order from other sites, or auto-post tracking from third-party purchases.

Shopify-Specific Rules

✓ Safe — Shopify specific

Use webhooks for real-time updates instead of polling the store. Better for Shopify's systems and keeps our data more current.

✕ Prohibited — Shopify specific

No redirecting away from Shopify Checkout. All transactions must go through Shopify's checkout system. We will never send customers to an external payment page.

Bottom line: We use official APIs on both platforms, keep a human (Chris) in the loop for all actions, and never scrape or automate purchases. This approach protects the 99.9% eBay feedback score, the Shopify store, and the business.

Phase 1: Automated Pricing Engine Up Next

How it works

When Chris lists a part (or reviews an existing listing), the pricing engine will:

Search eBay via Browse API for the same OEM part number. Returns active competitor listings with prices, conditions, and seller ratings.
Pull our own sales history from Supabase — what we've sold this part for on both eBay and Shopify, how fast, and at what margin.
Calculate suggested prices for both platforms — eBay price accounts for ~13% fees; Shopify price can be more competitive since fees are only ~3%. Both informed by the same market data.
Show Chris the recommendation with a confidence score and market context (how many competitors, price range, etc.).

Chris decides whether to accept, adjust, or ignore. The system never changes prices on its own without approval.

Data sources

eBay Browse API (production keys active) — searches active listings by keyword, part number, or category. Returns prices, conditions, seller info. Primary competitive intelligence tool.

Product Research in Seller Hub (manual, free) — 3 years of sold data including actual sale prices and sell-through rates. Seeds our pricing model with historical data.

Our own sales data from both channels — eBay orders (1,249 orders, $110.7K) and Shopify orders (898 orders, $87.4K) now unified in Supabase. Cross-channel price comparison reveals where we're leaving money on the table.

Phase 2: Agent-Based Inventory Management Planned

What the agents do

Lightweight AI agents that run on a Mac and communicate via WhatsApp or Telegram. Key agents:

Inventory Health Monitor: Daily scan of all listings across both platforms. Flags items listed 60+ days with no views (suggest markdown), items priced 20%+ below market (suggest increase), and fast-sellers that are low on stock.
Listing Assistant: Chris sends a photo + part number. The agent identifies the part, pulls fitment data, checks pricing, and drafts listings for both eBay and Shopify simultaneously.
Market Watch: Daily briefing on competitor price changes for our top SKUs, new listings in our categories, and trending parts.

Compliance note: All agents produce recommendations and drafts. Chris reviews and approves every action before it goes live on either platform. No autonomous listing, repricing, or purchasing.

Phase 3: Shopify Growth Planned

The big idea: eBay is our learning engine — it tells us what buyers want, what they pay, and what keywords they search. Shopify is our profit engine — every sale there saves ~10% in fees vs eBay. Phase 3 uses eBay intelligence to drive Shopify growth across three fronts: traffic migration, volume growth, and brand independence.

3A — Unified intelligence dataset

Both channels now feed into Supabase (2,147 combined orders). We can answer questions like:

Cross-channel pricing gaps: Which parts sell for more on eBay vs Shopify? Where should we adjust?
Channel-specific demand: Which parts sell only on eBay and need Shopify SEO to capture direct traffic?
Fee optimization: At $14.4K in eBay fees vs $2.8K in Shopify fees, which high-volume parts should we push harder on Shopify?

3B — SEO: six buyer search patterns

Auto parts buyers search in six distinct ways. Most competitors optimize for one or two:

Part number: "05143319AE" — OEM numbers in titles, URLs, meta, body text
Fitment: "2006 Jeep Grand Cherokee clock spring" — YMM in pages + structured data
Symptom: "steering wheel buttons not working" — blog content → part → product page
Category: "used OEM ABS module" — category pages with buying guides
Comparison: "OEM vs aftermarket clock spring" — compatibility guides + FAQ schema
Price/deal: "cheap Ford F-150 shift lever OEM" — pricing + Google Shopping feeds

3C — AEO: getting cited by AI assistants

When a buyer asks ChatGPT "Where can I find a 2006 Dodge Ram SKIM immobilizer?" — HubesHub should be the answer.

Structured data on every product page — schema.org with part numbers, fitment, pricing, availability
Write for fact extraction — specific OEM details, not generic marketing copy
FAQ schema — real buyer questions with direct, factual answers
Allow AI crawlers — GPTBot, ClaudeBot, PerplexityBot in robots.txt
Scale blog to 20+ articles — each diagnoses a problem and links to the solution

3D — Google Shopping & paid discovery

Start with Google Shopping ads for the top 20 highest-margin, fastest-selling parts
Use eBay sell-through data to identify proven demand
Budget: $5-10/day, measure ROAS, scale what works
Every Shopify sale saves ~10% vs eBay — paid traffic can still be profitable at higher CPA

3E — eBay-informed content pipeline

Pull top 50 search terms driving eBay traffic → create matching Shopify content
Priority: diagnostic/symptom guides first (highest buyer intent, lowest competition)
Vehicle-specific collections: "Everything we carry for 2004-2008 F-150"
Video content (YouTube/TikTok) driving traffic back to Shopify

Phase 4: Sourcing Pipeline & Shopping Lists Planned

The big idea: Jodi has built a working supply-demand pipeline — Python scrapers pull inventory from 5 local salvage yards (~15,000 vehicles), match against a master parts list (4,895 targets), and output weekly shopping lists. Phase 4 automates this pipeline, builds a real parts database, adds ROI-based ranking, and extends sourcing beyond the local yards.

4A — Automate Jodi's existing pipeline

Today: Jodi manually runs 5 Python scrapers and 5 PowerShell scripts from her machine. If she's busy, the lists don't get generated. The pipeline needs to run itself.

Current	Automated
Jodi runs scrapers manually	Scheduled cron (daily or 2x/week)
CSV files saved locally	Centralized in Supabase
PowerShell generates XLSX	Web-based shopping list on hubesiq.com
Emailed to Chris	Push notification when ready
No error alerting	Alerts if a scraper fails or a yard changes format

Technical approach:

Port Jodi's 5 Python scrapers into a unified runner with error handling and retry logic
Schedule via cron or a lightweight task runner on Mac (launchd) — no cloud infra needed initially
Write results to a yard_inventory table in Supabase with dedup (replace u_pullandsave_seen.txt with proper DB dedup)
Track vehicle arrival dates to distinguish "new this week" from "been sitting"
Build a simple status dashboard on hubesiq.com: last run time per yard, vehicle count, error log

Compliance note: These scrapers access salvage yard websites, NOT eBay or Shopify. Each yard's terms should be reviewed independently. This is Jodi's existing workflow being centralized, not a new scraping operation.

4B — Parts database (the demand signal)

The MasterPartsShoppingList.xlsx (4,895 rows) is the heart of sourcing, but it's a flat spreadsheet with shorthand abbreviations. It needs to become a proper database.

What we're building:

Normalized parts table — maps Jodi's shorthand (BCM, SKIM, TIPM, PS SB, RV Mirror Dim Onstar) to full part names, eBay listing titles, and Shopify product types
Fitment data — Year, Make, Model linked to each target part. Enables "which parts can we pull from this vehicle?" lookups at the yard
Market value per part — pulled from our own sales history (Section 08 revenue data) and eBay Browse API (Phase 1 pricing engine). Answers: "Is this part worth pulling today?"
Pull cost per part — seeded from PartPrices-USAuto.csv (~500 rows of yard pull prices). Extended to other yards as data becomes available
Sell-through velocity — how fast does this part type sell once listed? Derived from our order history + eBay Product Research data

Schema additions to Supabase:

target_parts — master list of parts we want, with full names, shorthand, fitment, category
yard_inventory — what's currently at each yard (vehicles + arrival dates)
yard_pricing — cost to pull each part type at each yard
part_market_value — current market value from eBay/Shopify data, updated with each pricing engine run

4C — Smart shopping lists with ROI ranking

The current lists answer: "Which yard vehicles have parts we want?" The smart lists answer: "Which parts should we pull first, ranked by how much money they'll make?"

ROI score per part:

Expected sale price (from pricing engine / sales history)
− Pull cost (from PartPrices-USAuto.csv or yard-specific pricing)
− Estimated fees (13% eBay or 3% Shopify, weighted by channel mix)
× Sell-through probability (from our history + eBay data)
= ROI score

Shopping list features:

ROI-ranked — highest-profit parts at the top, not just alphabetical
Already-in-stock flag — cross-references current eBay + Shopify inventory to avoid pulling duplicates
Vehicle row location — from yard data, so Chris/Luke can plan a route through the yard
"New this week" highlight — vehicles that arrived since the last visit get flagged
Web-based on hubesiq.com — mobile-friendly for use at the yard (replaces XLSX email attachments)
Per-yard view — filter by which yard Chris is visiting today

Role-based access (future, with Supabase Auth): Luke sees the shopping list for his assigned yard. Chris and Jodi see everything.

4D — Online sourcing beyond local yards

The 5 local yards are the primary supply chain, but they're geographically limited. Online sources expand the pool for high-value parts that aren't showing up locally.

Tier 1: Aggregator monitoring (highest value)

Car-Part.com — the industry standard for used OEM parts. 200,000+ recyclers. Has its own search interface and (unofficial) data feeds. Monitor for target parts at favorable prices.
LKQ Online — the largest auto recycler in North America. Inventory search available. Some API access for qualified resellers.
PartCycle — online marketplace for recycled OEM parts. Integrated inventory feeds from partner yards.

Tier 2: Wholesale and auction sources

Copart / IAAI — insurance auction vehicles. Buy whole vehicles for part-out at known cost basis. Higher upfront investment but better margins on high-demand parts.
eBay wholesale lots — sellers liquidating bulk auto parts. eBay Browse API can monitor these listings programmatically (fully within API terms since we're using the Browse API as a buyer, not scraping).

Tier 3: Direct yard partnerships

Some larger salvage yards offer CSV/API inventory feeds to partner resellers
Worth exploring with yards beyond the current 5 — especially yards in different regions with different vehicle mixes

How it connects: Online source prices feed into the same ROI calculation. If Car-Part.com has a SKIM module for $30 + shipping and we sell them for $98, that's a viable buy even with shipping costs — as long as sell-through is strong.

Compliance note: Car-Part.com, LKQ, Copart, etc. each have their own terms of service. Before building any integrations, we'll review each platform's terms for automated access. Some offer official APIs or data feeds; others may require manual monitoring initially. eBay Browse API monitoring of wholesale lots is fully within eBay API terms.

Phase 5: Scale & Competitive Moat Future

Long-term opportunities

Predictive demand modeling — anticipate which parts will be in demand based on vehicle age cohorts and seasonal patterns
Multi-channel expansion — add Amazon, Facebook Marketplace using the same unified inventory and data pipeline
VIN-based fitment search on Shopify — "Does this part fit my car?" as a customer-facing tool
YouTube/TikTok content pipeline — install videos and part guides driving traffic back to Shopify
Yard partnerships and geographic expansion — partner with yards in other regions for different vehicle mixes and higher-value finds

Questions for Chris

To build the best tools, we need Chris's input on these:

Pricing

How do you currently decide on a price? (Manual eBay search? Formula from cost? Gut feel?)
What margin do you typically target? (e.g., 40% over cost, 2x cost, varies by part?)
Do you price eBay and Shopify differently today, or same price on both?
At what point do you markdown a slow-moving part? (30 days? 60? Never?)

Inventory Workflow

When a part comes in, what's the flow? Photos → identify part number → list on eBay → then Shopify? Or batch?
Do you track cost per individual part, or per vehicle?
What's your monthly volume of new parts listed?
How does M2E Cloud sync work today, and what breaks?

What Would Help Most

What's the most tedious part of your current workflow?
If you could automate ONE thing tomorrow, what would it be?
What's your biggest concern about adding automation?

Business Goals

Is the primary goal to reduce eBay fees, increase volume, or both?
Are you open to investing in an eBay Store upgrade if the ROI is clear?
What does success look like for the Shopify store in 6 months?