We run, test, train and fine-tune open-source AI models locally on our own hardware, then build products and services on top of them that solve real-world problems.
Every project runs end to end, from the data to the people who use it.
Open-source models, run locally on our own GPU and measured on real tasks before we trust them.
Adapt the models that show promise to the language, data and problem in front of us.
Turn what works into products and services, with evaluation, safety checks and people in the loop.
What we are building, where each one stands, and what comes next.
Natural Sudanese Arabic voices from text. Used by thousands of creators, ad designers and podcast owners.
A generative image and video studio running open models on our own GPU.
Reading number plates and classifying vehicles from motorway video, and checking uncertain reads before anyone is charged.
Plans, progress and lessons, as they happen.
Zol AI Lab is now the home for our research and products. We run, test, train and fine-tune open-source AI models locally, then build products and services on top of them that solve real-world problems.
When a camera is unsure whether a character is a 1, an I or a T, the wrong person can be billed. Our verifier lists the likely spellings, looks each one up in the registration record, and compares the registered vehicle with what the camera actually sees: body type, make and badge, described by a local vision-language model that never sees the plate.
The first version of ZolStudio runs image, video and lip-sync models on our own GPU, behind one simple studio: pictures from words, video with sound, talking shots from a portrait and a voice, and one-click character swaps in existing clips.
ZolVoice, our Sudanese Arabic text-to-speech product, is live at zolvoice.com. It routes each script to the right voice model, fixes Sudanese pronunciation before speaking, and answers support emails with an AI agent that a person approves.
Zol