This project is building a practical Dholuo AI system that can help a Dholuo speaker be understood by a non-Dholuo speaker and viceversa through real-time AI translation and voice technology. Your support helps fund paid Luo contributors, data collection, model training, platform maintenance, and public progress reporting.
Pilot goal: first public Dholuo AI demo with speech understanding, translation, and voice output.
Many African languages are still missing from mainstream AI tools. This project focuses on Dholuo so speakers can be better understood in digital spaces, while also creating a fair community-driven process that pays Luo contributors to help build the language data the model needs.
Dholuo should not be invisible online. Supporting this work helps preserve language, identity, and practical everyday use in the AI era.
The project creates paid work for Luo speakers through transcription, voice review, annotation, and language quality checks.
The goal is not just research. It is a usable system for voice understanding, translation, and spoken Dholuo output.
When you support this project, you are helping create something practical and long-term for Dholuo speakers. Your contribution helps turn an important idea into a working tool that improves communication, creates paid work, and gives Dholuo a stronger place in technology.
Build a system that helps Dholuo speakers communicate more easily with non-Dholuo speakers through speech and text translation workflows.
Create a model that can respond in spoken Dholuo so users hear natural language output instead of text alone.
Open opportunities for Luo contributors to earn through transcription, validation, and voice data preparation.
This project is being built step by step so you can easily understand where the work starts and where your contribution goes. From collecting language data to training the model and releasing a public demo, each stage is designed to show visible progress and real impact.
As a supporter or visitor, you should be able to see that the work is moving and that the project is staying accountable. This section is here to show clear milestones, visible targets, and measurable progress so you know what your support is helping achieve.
Early supporters interested in the mission
Pilot transcription and review target
Luo contributors for data tasks and QA
Pilot fundraising target for first release
This is where you can quickly see how far the campaign has gone and how much is still needed to reach the first release. As more people contribute, this progress area should give every donor confidence that the project is active, transparent, and moving toward a real outcome.
KES 790,000 raised so far — approximately 33% of target.
Before you contribute, you should be able to see exactly what your support will fund. This pilot budget is designed to cover the practical work required to build a credible first release of Dholuo AI, including paid contributor work, speech and text data preparation, model training, technical infrastructure, and transparent reporting.
| Item | Estimated cost | Purpose |
|---|---|---|
| Transcription and annotation | KES 780,000 | Pay trained contributors to transcribe, segment, label, and clean Dholuo text and audio data for model training. |
| Voice recording stipends | KES 320,000 | Support speaker recruitment and compensate contributors for recording clear, usable Dholuo speech samples across different voices and accents. |
| Language review and QA | KES 210,000 | Cover second-pass review, consistency checks, correction of transcripts, and quality control before data enters training workflows. |
| Model training and compute | KES 460,000 | Fund cloud compute, model experiments, storage, inference testing, and the technical work needed to produce a usable pilot system. |
| Platform engineering and hosting | KES 230,000 | Maintain the website, support contribution workflows, backups, uptime, security updates, and public-facing project pages. |
| Contributor onboarding and support | KES 140,000 | Prepare contributor guides, task instructions, onboarding calls, support materials, and day-to-day coordination for distributed workers. |
| Operations, payments, and reporting | KES 160,000 | Handle mobile money disbursements, reconciliation, admin follow-up, public updates, and milestone reporting to supporters. |
| Compliance and contingency | KES 100,000 | Provide a buffer for unexpected technical needs, revisions, contributor replacements, and essential project administration. |
| Total pilot estimate | KES 2,400,000 | From paid language data work and voice collection to model training, infrastructure, and public pilot release. |
Your support matters, and you deserve the confidence that it is going to a real, accountable project. This section helps you see who is behind the work, how contributions are handled, and what practical safeguards are in place so you can give knowing your support is helping move Dholuo AI forward in a transparent and credible way.
Project lead: Douglas Assesa
Role: Product lead and project coordinator for Dholuo AI.
Focus: Building practical Dholuo speech, translation, and language data systems that create both digital access and paid local work.
Email: hello@dholuo.digital
Phone / WhatsApp: +254 734 220 002
Location: Nairobi, Kenya
Working model: Small independent project team working with paid Dholuo contributors, reviewers, and technical collaborators.
Supporter updates: Monthly milestone summaries published on the project site as fundraising and delivery progress continue.
This campaign is structured to show who is responsible, what support is funding, and how progress will be communicated as the pilot moves toward public release.
If you are thinking about supporting this project, it is normal to want clear answers first. This section is here to answer the most common questions about the purpose of the project, how contributions are used, how people will be paid, and how progress will be shared publicly.
Because Dholuo speakers should be understood in digital spaces, customer support, health, education, media, and everyday communication. This project helps bring Dholuo into modern AI systems through speech, text, and translation tools.
Your contribution supports paid transcription work, voice recording, quality review, model training, contributor onboarding, website maintenance, and public progress reporting.
Yes. A key part of the project is creating paid online work for Luo contributors who transcribe, review, and validate Dholuo language data.
The first release aims to understand spoken Dholuo, translate key Dholuo phrases into English or Kiswahili, and speak back in Dholuo for selected real-world use cases.
The platform should publish regular updates showing money raised, work completed, data collected, contributors onboarded, and milestones achieved.
Your contribution helps build a practical AI system for Dholuo, create paid work for Luo contributors, support transcription and training, and keep the platform running with clear public updates.