This article answers the most common questions about Smartling connectors and plugins, including how to choose between connectors and the GDN, how to configure projects and environments, what visual context is available, and how translation edits are delivered back to connected platforms. Use the links below to jump to the section on a particular topic. You can also see the full list of Smartling connectors and plugins.

For an in-depth look at how hosted connectors upload content, detect changes, and deliver translations, see Overview of Hosted Connector Behavior.

Connectors vs. GDN: Choosing the right solution

Because the right approach depends on your content platform, site architecture, and workflows, we recommend working with a Solution Architect to determine which solution best fits your setup. At a high level, a connector can be a good fit when your content lives in a specific platform, such as a CMS, support platform, or design tool, while the Global Delivery Network (GDN) can localize a website or web application regardless of its underlying technology. Both solutions let you translate content without manually extracting source strings.

Connectors establish a direct integration with a specific content platform, such as a CMS, support platform, or design tool. If you already work within one of these platforms, a connector (particularly a hosted connector where translations are requested and managed within the Smartling dashboard UI) is often the most straightforward option.

Connectors also handle regionalized content well. Many CMS platforms support region-specific source content, where the source itself varies by region or market. Because connectors integrate directly with the platform, they can treat each regional variant as distinct source content and manage it independently. This level of regional content management is difficult to achieve with the GDN.

The GDN can localize any website or web application regardless of the underlying technology. It works by intercepting and translating content in transit between the web server and the browser, so it does not require a direct integration with your content platform.

Other factors to evaluate include performance at scale, ongoing maintenance overhead, and pricing. See Introduction to the GDN for more details on how the GDN works.

Can one Smartling project connect to multiple environments, and vice versa?

We strongly recommend maintaining a 1:1 relationship: one environment in the connected platform linked to one Smartling project. For most hosted connectors, this is the only supported configuration.

While it is technically possible to connect one environment to multiple Smartling projects, this is generally not recommended unless each project handles a distinct subset of content. If two projects translate overlapping content, they must be configured for different target languages. Otherwise, the projects can overwrite each other's translations.

There is no charge based on the number of Smartling projects you create, so you are free to create as many as you need. Each project is associated with a linguistic package (its translation memories, style guides, and glossaries), which is typically aligned to a specific content type (for example, Marketing, Internal, Legal, or UGC). Because the same linguistic package and assets can be shared across multiple projects, translations completed in one project can be leveraged by another through SmartMatch when both projects use the same translation memory.

What visual context is available for my connector?

Visual context provides translators with a visual representation of the source content while they work in the CAT Tool. The type and quality of visual context available depends on your connector. A connector's ability to generate visual context largely depends on whether the connected platform offers a preview API that the connector can access. This is most common with headless CMSs, where content creation is separated from the logic that renders the final user experience. To generate high-quality visual context, the platform must support an API-driven preview of the rendered content so the connector can display what an end user would actually see.

Default visual context

Nearly all hosted connectors provide default visual context, which is automatically generated from the source content. This is a simplified HTML representation of an asset's fields and values. It presents the content in a basic layout but does not reflect how it will appear when styled or formatted in the live application.

Visual context from content file

If a connector sends files to Smartling in a native format that supports visual context generation (for example, the Google Drive connector sending Docs, Sheets, or Slides), visual context is generated automatically from those files. Design application plugins (Figma, Adobe InDesign, Illustrator, and Photoshop) also provide visual context automatically by sending image snapshots of the design files.

Example context generated from a Google Slides presentation:

Integrated visual context

Some connected platforms offer a preview API that the connector uses to render a preview of how the content will appear to an end user.

Example context generated using the integrated context option for the Contentful (Fields) Connector:

Built-in visual context capture

Some non-hosted connectors, including AEM Touch, AEM Cloud, the Sitecore XP Connector, and the Drupal TMGMT Connector, have built-in visual context capture tools that are triggered when a translation request is submitted. The connector captures a rendered page preview and uses it to generate visual context.

Connectors with no visual context

The GitHub Connector and Repository Connector are designed for translating developer resource files (JSON, YAML, PO, and similar formats). These files contain source strings only, with no information about how content is rendered in the application. For these connectors, visual context can be added via the Smartling API or the JavaScript Context Capture Library.

If you have questions or need guidance on generating visual context for your connector, work with your Solution Architect or contact another Smartling representative to discuss available options.

My job is completed, but I don't see the translations.

A completed job does not always mean an asset is fully translated and delivered. Most connectors are file-based: they track translation progress at the file level, not the job level, and an asset's strings can be spread across more than one job. This commonly happens when the source asset is updated later on, and the new or updated content is translated as part of a separate job from the original content. Translations are delivered only once every authorized string for a file is complete for a given target locale, so if any string for that locale is still in progress or awaiting authorization, delivery does not trigger, even when a job shows as completed.

For hosted connectors, the step-by-step way to diagnose this (checking the asset details, the translation progress bar, and the Strings View to find which jobs hold the incomplete strings) is documented in Overview of Hosted Connector Behavior.

How do I ensure edits made to published translations are pushed back to the connected platform?

Always make edits in Smartling, then ensure delivery is triggered for your specific connector. Most hosted connectors automatically re-deliver edited translations. The main exception is the GitHub Connector, which is job-based rather than file-based, so edits made after a job completes are not sent automatically. For non-hosted connectors, behavior varies: some deliver updated translations automatically during their next scheduled cron run, while others require an additional step. The table below summarizes the behavior for each connector.

Always edit translations in Smartling

All translation edits must be made in Smartling, not in the connected platform. The connector does not recognize changes made directly in the connected platform, and those changes may be overwritten the next time the connector delivers translations.

Editing in Smartling is also critical for maintaining translation memory (TM) quality. When you edit a published translation in Smartling, the corrected translation is saved to the TM and can be used for future leverage and SmartMatch. Edits made outside Smartling bypass this process, which means incorrect translations may be reused in the future.

What counts as editing a translation in Smartling

"Editing a translation" means changing the translation of a project string, which is the translated string associated with a specific file in your project. You can do this in any of the following ways:

• Edit the string in the Strings View using Quick Edit.
• Edit the string in the CAT Tool and save.
• Edit the translation in the translation memory and make sure the change is applied to the project strings: choose Save & Update Projects when editing a single unit, or select the Update Project Strings checkbox when using Find and Replace.

All of these methods update the project string. Updating only a TM entry without applying the change to the project strings is not enough to trigger delivery of the updated translation. Because connectors work with files, and project strings belong to specific files, a connector only delivers a translated file when the project string in that file has changed. When editing in the TM, you must apply the change to the project strings (using Save & Update Projects or the Update Project Strings checkbox in Find and Replace) rather than updating the TM entry alone.

For an overview of the available editing methods, see Options for Editing Translations.

Delivery behavior by connector

Once you have made edits in Smartling, how those edits get pushed back to your connected platform depends on which connector you are using. The table below summarizes the behavior for each connector type.

Zendesk Support and Salesforce Service Cloud use instant translation and do not ingest content into the Smartling dashboard for translation. Content does not go through a job or workflow, so the editing and re-delivery process described above does not apply to these connectors. Consult the documentation for each connector for details on how translations are managed.

Is prepublishing supported for all connectors?

Prepublishing allows you to deliver in-progress translations to your connected platform before the strings in the file reach the final published step in the workflow. Generally it should only be enabled if you have a specific business need, as it can add unnecessary load to the integration and carries the risk that users may access unreviewed translations.

Hosted connectors

Prepublishing is supported for all hosted connectors. However, prepublished translations have not yet reached the final step of the workflow and may not have completed all review and quality assurance steps. For this reason, prepublishing should only be enabled when there is a specific business need to access translations before they are fully published.

For example, if prepublishing is enabled for the Google Drive Connector, translated documents may be created in Google Drive before the translation job is complete. Depending on your sharing settings, users could have access to translations that are still undergoing review.

If you are using the GitHub Connector, prepublishing is supported through the Early Delivery feature. See Get Translations Before Job Completion with Early Delivery.

If you need prepublished translations occasionally

If you only need access to in-progress translations occasionally, you do not need to enable prepublishing. Instead, you can manually export translations for a specific asset directly from Smartling. Simply locate the asset and select Export Translations either from the Actions menu or from within the asset details page.

Export from Actions menu:

Export from asset details page:

If you need prepublished translations on an ongoing basis

Contact your Solution Architect or another Smartling representative to enable it for your project. Enabling prepublishing requires enabling a special callback for prepublished translations, so the connector delivers translations when all authorized strings in an asset are prepublished.

How to enable prepublishing for a hosted connector:

1. Contact your Smartling Solution Architect or another Smartling representative to enable the "Callback on prepublish" callback for your project and change the retrieval type to "pending" within your project's Smartling settings. (Only accessible to Smartling admins, so you will not see these settings in your project.)
2. Ensure you have enabled prepublishing for a workflow step. For example, enabling prepublishing on the translation step of your workflow so translations are prepublished as soon as a translation is submitted in that step.
3. Authorize content for translation using that workflow. The connector will deliver translations once all authorized strings from an asset are prepublished, and again once all the authorized strings are published.

For more information on prepublishing, see Prepublish Translations.

Non-hosted connectors

Support for prepublishing varies by connector. Refer to the documentation for your specific connector or contact a Smartling representative to confirm whether prepublishing is available.

Design plugins

For Smartling design plugins, including Figma, Adobe InDesign, Adobe Illustrator, and Adobe Photoshop, you can retrieve and apply prepublished translations before a translation job is complete. However, this approach is generally not recommended because the translated design file will need to be updated again once the final translations are published, resulting in multiple versions of the translated file.

In most cases, it is best to wait until translations are fully published before applying them to your design files.

How do connector versions and upgrades work?

Hosted connectors and design application plugins (Figma, Adobe InDesign, Illustrator, Photoshop): Updates are pushed automatically by Smartling. No action is required on your end.

Non-hosted connectors: You are responsible for upgrading to the latest version. The upgrade process varies by connector. Refer to the documentation for your specific connector for instructions.

To stay informed about new connector releases, join the dedicated space for your connector in the Smartling Community under "Integrations & Developer Tools" and enable email notifications.