mondialrelay_relay

Some carriers deliver to a pickup point rather than to an address, and the customer has to choose which one. That choice is a map, and the carrier publishes their own widget for it, because only they know where their points are and whether each one is open.

Rather than reimplementing that, Odoo embeds it. This widget loads the carrier's script from their servers, renders their picker into a container on the delivery method dialog, and configures it from the record: the brand, the service mode, which countries are allowed, and the customer's postcode and country so the map opens somewhere useful.

When the customer picks a point, the carrier's script hands back its details, and the widget stores a JSON summary of the address in the field. So what Odoo keeps is a snapshot of the chosen point rather than a link to a record.

Pickup point delivery is cheaper than to-the-door and, in several markets, what customers prefer. Making the choice inside Odoo rather than on the carrier's site is what makes it usable for an internal sales or support user preparing an order on the customer's behalf.

The dependency is the flip side. The map is served from the carrier's infrastructure, so it needs the user's browser to reach the internet, and an outage or a change on their side shows up as a broken area inside your ERP. This widget is the clearest example in core of that trade, and it is worth naming explicitly in any project that relies on it.

The bigger planning point is that the module is gone in the next version. Whatever replaces it, a database that depends on this integration should not be upgraded without checking what is available first.

The core usage

<field name="mondialrelay_last_selected"
       widget="mondialrelay_relay"/>

No options. The widget renders nothing at all unless the chosen delivery method is that carrier.

What it writes

{
  "id": "FR-012345",
  "name": "Point Relais",
  "street": "...", "zip": "...",
  "city": "...", "country": "FR"
}

A JSON snapshot of the chosen point, stored as text on the record.

The values it configures the map with

is_mondialrelay          # render or not
mondialrelay_brand       # account identifier
mondialrelay_allowed_countries
shipping_zip, shipping_country_code
mondialrelay_last_selected_id  # pre-selects a point

All read from the record, none declared as a dependency.

Three things make this widget unusual in Odoo core, and all three are worth knowing before depending on it.

It loads a third-party script by URL at render time, guarded so the load happens once. That is rare in core, and it makes the widget's availability a function of the user's network rather than of the database.

It talks to that script through the jQuery plugin interface the carrier publishes, wrapping its container element and calling their initializer with a parameters object, then triggering a rebind so the map lays out correctly inside a dialog.

And it contains a workaround written against a bug on the carrier's side. When the picker reports no results, a comment explains that the carrier's script can crash with a cross-origin error, and the widget temporarily suppresses Odoo's third-party script error handler so that crash does not surface as an Odoo error dialog. It is candid about being a hack, and it is a good example of the kind of maintenance an embedded third-party widget brings with it.

The rendering guard is simpler but important: the container element only exists when the record says the chosen method is that carrier, so on any other delivery method the widget renders nothing and loads nothing.

Odoo 20 is expected at Odoo Experience in Brussels, 24 to 26 September 2026. The development branch is unstable and this may still change; we re-verify the page after release.

The module is gone. The carrier integration module is not present on the development branch, and neither is this widget. That is consistent with a wider reshaping of which carrier integrations ship in the core repository.

For a database using this integration, that is a question to answer before upgrading rather than a migration to perform: what replaces it, and whether the pickup point data you have stored still means anything to the replacement.