The Lab
Some things ship. Some get archived. All of it teaches.
dbar Labs is an applied AI lab as well as an implementation studio. Experiments get built, tested in public, and kept only if they earn their place. When something does not work, it gets retired, and the write-up stays on this page.
Everything below carries an honest status. Nothing dead is dressed up as alive, and nothing early is described as proven.
How to read the tags
Public demonstration project. Not a commercial offer.
Archived experiment. Kept for the record, not maintained.
dbar Hybrid
A static public demonstration that turns a fictionalized 32-week HYROX-to-marathon sample plan into a clear, mobile-first training experience.
- Status since
- 2025
- Type
- Public demonstration
The problem
Training for HYROX and hybrid races is hard to plan. Generic programs do not adapt to your fitness, your schedule, or the week you actually had. Most athletes end up stitching together plans from five different places and guessing.
Who it is for
Everyday athletes training for their first or next hybrid race, not elites with coaches.
Where it stands
dbar Hybrid began as an AI-assisted training experiment developed from Daniel's own hybrid-racing experience. A small group of testers, including trainers, reviewed early versions, but it was never formally launched or commercialized. Active product development was discontinued. The current public demo uses a fictionalized sample athlete, requires no account, and demonstrates the product work dbar Labs can create.
What we learned
People do not want a perfect plan. They want a plan that adapts when life happens, explained simply enough to follow on a tired Tuesday.
dbarGPT
A document-grounded AI assistant for small businesses. Built in the summer of 2025, archived the same year.
- Built
- Summer 2025
- Archived
- 2025
The problem
Teams could not get straight answers out of their own documents. dbarGPT let them upload documents, ask questions, and get answers grounded in their own material.
Who it is for
Small businesses whose answers were trapped in PDFs, manuals, and shared drives.
Where it stands
dbarGPT was built during the summer of 2025 as a document-grounded AI assistant for small businesses. Friends who owned businesses tested early versions, but it never had paying customers or a formal commercial launch. Daniel archived the project because the underlying AI platforms were changing faster than a solo product could reasonably maintain. The experiment helped shape dbar Labs toward focused implementations instead of asking clients to adopt another standalone platform.
What we learned
Sell the outcome, not the tool. That lesson shapes every engagement dbar Labs takes on today.
How the lab works
Honest experiments, small scope.
01
Real problems first.
Every product starts from a problem Daniel has personally watched or lived. No trend chasing.
02
Small scope, shipped.
Products start narrow and earn their roadmap. dbarGPT taught that one.
03
Honest status.
Public demonstration or archived. The tag says where things actually stand, and it moves when reality does.
04
Human oversight.
A person is accountable for anything these systems say or do. That does not change as they grow.
The rule of the lab
Launch honestly. Test honestly. Retire honestly. If a product on this page ever claims traction, it will come with numbers we can show.
The next experiment could be your workflow.
The audit is how the lab studies a new business. Yours could be next.
Free · 15 minutes · English or Español · No obligation