Sync MercadoLibre Stock With Your Store: Complete Guide

Selling on MercadoLibre, Tiendanube and your physical store at the same time is great for revenue and a nightmare for stock. Syncing stock between MercadoLibre and your store is done with ML's official API: webhooks that announce every sale in real time, a central stock that deducts the unit, and automatic updates to the other channels in seconds. The same circuit works for prices. Without that sync, overselling isn't a risk: it's only a matter of time.

The real problem: the unit sold twice

The scenario we hear constantly: one unit was left, it sold in the store at 11:00, and at 11:20 a MercadoLibre order came in for that same unit. Now you have to cancel, and on ML cancelling has a price:

• Reputation: cancellations for lack of stock hit the thermometer directly. Going from green to yellow tier can cost you 20-40% of sales during the weeks it takes to recover it.
• Ranking: ML's algorithm penalizes listings from sellers with cancellations.
• Operational cost: every oversell means phone calls, apologies and sometimes a discount coupon to calm the buyer down.

And the mirror problem: outdated prices. You adjusted prices in your own store after a supplier increase, but ML kept the old price. You sold, sure — at a negative margin.

How the MercadoLibre API works (the essentials)

MercadoLibre has a mature, well-documented REST API. For syncing, there are four pieces:

1. OAuth 2.0: your application is authorized against the seller's account and obtains tokens (which expire every 6 hours — your system renews them on its own with the refresh token).
2. Notifications (webhooks): you subscribe to topics like orders_v2 (sales), items (listing changes), questions (questions) and claims (claims). When something happens, ML sends a POST to your server with the resource ID. Important: the notification doesn't carry the full data, it carries the pointer; your system then queries the detail. If the concept is new to you, here we explain what a webhook is and what it's for.
3. Item resources: with a PUT to /items/{id} you update available_quantity and price. If you use variations (sizes, colors), stock is managed per variation.
4. Usage limits: the API has rate limits; a serious syncer batches updates and handles 429 errors with retries, instead of bombarding it.

Typical multichannel sync architecture

The design we implement in most projects:

• Single central stock (in your ERP, management system or your own database): it's the single source of truth. The channels are mirrors.
• Inbound via webhooks: sale on ML → notification → your system deducts from central stock → propagates the new quantity to Tiendanube/Shopify and to the other ML listings sharing that SKU. Sale on Tiendanube → same circuit in reverse.
• Periodic reconciliation: in addition to webhooks, a process every 15-60 minutes compares published stock against the central one and corrects differences. Webhooks sometimes get lost; reconciliation is the safety belt (the logic of combining both is explained in cron jobs vs webhooks).
• Update queue: if ML is down or returns an error, the update is queued and retried. Nothing is ever lost.

With this architecture, the overselling window drops from hours to under 30 seconds, which in practice eliminates it except in launches with extreme demand.

One detail that trips a lot of people up: if you use Mercado Envíos Full, the stock sitting in ML's warehouses is managed separately from yours. The syncer has to distinguish between your own stock (which you control) and Full stock (which ML controls), and decide which listings each one feeds. Ignoring that distinction is the number one cause of homemade syncers that "worked fine until we turned on Full."

Are you cancelling sales for lack of stock or selling at old prices? Book a 30-minute meeting and we'll look together at how much it's costing you and how to fix it.

Automatic repricing: prices that defend themselves

The second layer of value. In ML's catalog you compete against other sellers for the same listing, and the price decides who wins the sale. An automatic repricer:

1. Checks competitors' prices on your catalog listings every 1-4 hours.
2. Applies your rules: "lower until you win, but never below cost + 18% margin", "if I'm the only one, raise to the defined maximum price".
3. Updates via API and logs every change so you can audit it.

Typical result in accounts with 200+ competed listings: 5-12% more sales won in the catalog without touching the minimum margin. The key isn't the algorithm, it's that the rules reflect your real cost structure — including ML commissions, shipping cost and taxes, which is where generic repricers fail.

Off-the-shelf solutions vs custom development

Before building, look at what already exists:

Option	2026 monthly cost	Ideal for	Limitation
Native Tiendanube↔ML connector	Included / low	Small catalog, 1:1	Simple rules, no ERP
Real Trends, Astroselling and similar	USD 30-150/month	ML-centric sellers	Predefined logic, cost grows with volume
WooCommerce/Shopify plugins for ML	USD 10-50/month	Own store + basic ML	Spotty support, conflicts between plugins
Custom development	USD 3,000-10,000 one-time	Own ERP/warehouse, kits, multi-account, per-channel rules	Higher upfront investment

The honest rule: if your operation fits the mold of an off-the-shelf tool, use it. Custom development is justified when one of these shows up: stock in an ERP or management system that no off-the-shelf tool integrates well, kits and bundles (selling SKU A deducts SKUs B and C), multiple ML accounts, different prices per channel with your own rules, or volume where the SaaS monthly cost exceeds the installment of your own build.

A home-appliance SMB we worked with went from 6-8 monthly cancellations for overselling to zero in the first month of custom sync, and recovered its green tier by the second. The system paid for itself before the first quarter.

What else you can automate while you're at it

The integration with ML's API opens the door to more than stock and prices:

• Automatic invoicing of every sale: the ML order triggers issuing the receipt with CAE — we explain it step by step in how to automate invoicing with ARCA.
• Answers to frequent questions with AI, within our approach to AI automation.
• Shipping tracking and notifications to the customer with no human intervention — a topic we cover in logistics shipping-tracking automation.

When you DON'T need automatic syncing

• You sell fewer than 30 orders per month in total: updating by hand with a daily 15-minute routine is enough. Re-evaluate when you double your volume.
• You have deep stock of everything: if you never run out of units, overselling isn't your risk (outdated prices maybe are).
• Your whole catalog changes every week (clearance, single batch): the cost of constantly mapping SKUs can exceed the benefit; there it's better to rethink the loading model before the sync.

The next step

If you've already been through the unit-sold-twice situation, you know the real cost isn't the cancelled sale: it's the reputation that takes you months to rebuild. At Deepyze we design and implement integrations with the MercadoLibre API and API development for multichannel sellers in Argentina and LATAM: central stock, repricing with your rules and automatic invoicing, all monitored in production. Fixed price, a proposal in 24 hours and a team in your time zone that knows what a Hot Sale is. Tell us about your operation and we'll tell you whether something off-the-shelf is enough or what it would cost to do it right, custom-built.