SettleIn US
Using Drupal AI to make critical refugee resources more accessible
Client Overview
When refugees arrive in the United States, they face an overwhelming amount of new information. Employment, housing, healthcare, immigration status, legal rights, and daily life in a new country all require immediate attention.
SettleIn US, a digital companion created by the International Rescue Committee (IRC), helps refugees navigate that transition. Available in 11 languages, the platform provides trusted information designed to support successful resettlement.
Services provided
SettleIn Mobile Gallery
The challenge
SettleIn US serves users arriving in the US with varying levels of literacy, language proficiency, and access to technology. Many users are navigating stressful life transitions and may not have the time, ability, or environment needed to read lengthy articles. A refugee parent doing intake paperwork after a 14-hour travel day needs another mode — voice — to absorb the same critical information a reader gets from the page. Producing professionally narrated audio in-house, for every article, every time content changes, in every language, was not realistic on the project’s budget or timeline.
To make that information more accessible, the IRC partnered with Kanopi to explore how artificial intelligence could help deliver content in a format that works for more people: audio.
While audio versions of content could improve accessibility, producing and maintaining narrated recordings across a large multilingual content library presented significant challenges:
- Recording professional narration for every article would be costly and time-intensive.
- Content changes would require re-recording and republishing audio.
- Supporting multiple languages would increase production complexity.
- The editorial team needed a solution that fit naturally into their existing Drupal workflow.
The organization needed a scalable way to provide audio content without creating a separate production process.
The process
Producing narrated content
The AI solution automates the creation of AI-generated audio files (mp3) that represent the text on the page. Content editors are empowered to generate the AI audio on individual posts and their translations as needed.
Overcoming barriers to unsupported languages
To overcome any limitations for languages that are unsupported, we implemented a field where the client can upload manually produced recordings in the unsupported languages (this can also be used for the supported languages when a different voice is desired).
The solution
Kanopi implemented an AI-powered text-to-speech (TTS) workflow within the site’s Drupal 11 platform.
Using Drupal AI and the ElevenLabs integration, editors can generate narration directly from Drupal whenever they choose. Rather than generating audio for the entire site, the solution gives editors control over which articles receive narration, helping the organization prioritize content and manage usage costs.
The workflow is simple:
- An editor enables AI narration on a piece of content.
- Drupal automatically assembles narration-ready text from the article.
- The content is sent to ElevenLabs for text-to-speech generation.
- The generated MP3 is attached to the article and displayed through an embedded audio player.
The solution currently supports English content and supported translations through ElevenLabs. For languages not currently supported by the service, Kanopi implemented a complementary workflow that allows editors to upload manually produced recordings, ensuring accessibility remains available across the site’s diverse language offerings.
Why Drupal AI?
Rather than building a custom AI platform from scratch, Kanopi leveraged Drupal’s growing AI ecosystem to create a solution that could be maintained and expanded over time.
The implementation combines:
- Drupal AI
- ElevenLabs
- Key
- Token
- Custom Drupal integrations for content preparation and workflow automation
This approach allowed the team to integrate AI directly into the editorial experience while maintaining flexibility for future enhancements.
The result
Before this project, SettleIn US did not provide narrated versions of its content.
Now, every published article on SettleIn US that opts in now ships with high-quality, consistent narration the day it’s published — no studio booking, no voice talent scheduling, no manual re-recording when content updates. Editors can create audio versions of articles directly within Drupal, gaining back the time they used to lose chasing audio production, and refugees get the same critical information in a format they can actually use.
The benefits extend beyond efficiency:
- Improved accessibility for users who prefer or require audio content.
- Reduced effort required to create and maintain narrated resources.
- Editorial control over when and where AI-generated audio is used.
- A sustainable workflow that scales with ongoing content updates.
Most importantly, the solution helps ensure that important information can reach people in the format that works best for them.
For an organization whose audience is, by definition, time-poor and under stress, removing the friction between “article published” and “article listenable” is the win. Drupal AI made it a one-checkbox feature.
Looking Ahead
As AI language and voice technologies continue to evolve, SettleIn US now has a foundation for expanding audio accessibility across additional content and languages.
By combining Drupal’s flexible content platform with practical AI tools, the organization has created a scalable approach to making critical information easier to access for the people who need it most.
