I am a language service provider (not a dev) using the cloud-based translation management system Memsource.

Memsource has multiple connectors and integrations that allow it to talk directly to other systems (Dropbox, Google Drive, Zendesk, Salesforce, Atlassian). There are also plug-ins for WordPress and other CMS.

Content can be read from these systems and, once translated, written back without manual file transfers. E.g., an English Zendesk article can be read from a Zendesk site, translated to French in Memsource, and the French translation will be written back to the Zendesk site as a separate article.

https://www.memsource.com/integrations/

One of my clients runs multiple translated websites using Umbraco. Here, each language is a separate “branch” in their Umbraco installation mapped to TLDs. Example:

• root
  • en
  • products
    • product X
    • version X1
  • fr
  • produits
    • produit X
    • version X1

So the English source node “product X” (with subnodes) is copied (manually) to the “/fr/” branch, where the English content is overwritten with a French translation. It then gets published as company.fr/produits/produit_x/

(I realise this can be done differently. But this is the setup we have, and it won’t change.)

Unfortunately, there is no Memsource connector for Umbraco. So to benefit from Memsource’s features (translation memory, machine translation etc.), I currently have to copy translatable content from an Umbraco node to plain text / HTML files, which are processed in Memsource, then pasted back into the target branches (/fr/, /de/ etc). This is obviously a serious waste of time.

I am looking for someone who might be able to write such a connector using the Umbraco and Memsource APIs.

https://help.memsource.com/hc/en-us/articles/115003694732-Introduction-to-Memsource-API

https://cloud.memsource.com/web/docs/api

Here is how it should work:

• The connector is added to the existing Memsource (Team) account.
• A translation project is created in Memsource.
• A translation job is created in the project.
• Using “Add from Online Repository”, the client’s Umbraco site is accessed.
• A page / node (and its subnodes) can be selected for translation.
• Content is read / extracted from the selected node and presented in Memsource for translation.
• Translation is done in Memsource (this is the core / standard functionality).
• Once translation is finished, the translated content can be written back to the Umbraco site by selecting a target node (e.g. in the “/fr/” or “/es/” branches) that will be overwritten with the translated content.

What makes this tricky is that the Umbraco site in question is heavily customised. Templates have plain text and rich fields. All this needs to be extracted and read into Memsource so that segmentation (splitting content blocks into sentences) works as with all other formats.

I thought about approaching this as follows:

1. Set up a free Memsource Developer account
2. Set up an Umbraco test / sandbox system.
3. Implement the functionality described above.
4. Once it works, coordinate with the client’s Umbraco agency to integrate the connector into the client’s site.

There’s a (somewhat modest…) budget for this project.

Does anyone know of an existing solution? Has someone here done this or would be interested in developing it?

Thank you.