The problem
Specialist reference information is often scattered across sources. People need a clear way to find a cartridge, understand the source, and compare information without losing the caveats.
Reference & decision tools
Caliber Atlas
Specialist information, made easier to explore.
- Built for
- People exploring specialist reference information
- Our work
- Product design, development & source presentation
- Current focus
- Useful tools with inspectable methods and limits
What we’re building
Caliber Atlas brings a searchable cartridge reference, comparison tools, and educational resources into one place. It distinguishes manufacturer-published inputs from modeled results and keeps source information close to the data.
Where AI fits
AI-assisted research and software development support how we build and refine the product. Reference claims still need named sources; software behavior still needs testing. The site is an educational reference with its own published methods and limits.
What this work demonstrates
Searchable knowledge tools with visible sources and understandable limits. The transferable idea is simple: help a reader distinguish what is published, what is estimated, and what still needs checking.
A CLOSER LOOK / DESIGNING FOR TRUST
Make the evidence
part of the experience.
A reference is more useful when a reader can see where the information came from and what still needs checking.
What the source says.
Manufacturer-published information, with its source and context.
What the model estimates.
Calculated results, with assumptions and limits kept visible.
What someone observed.
User-recorded observations kept distinct from published references.
The design decision
Give published claims, document reviews, stored entries, and modeled results distinct labels. Each label describes the kind of evidence behind the information.
The check readers can make
The public methodology explains how source reviews work and where their coverage is incomplete. Readers can inspect the process and submit corrections with supporting sources.
The limit we keep visible
A reviewed document is not an independent physical measurement. Evidence gaps and model limitations remain part of the product.
YOUR NEXT IDEA
What could work
better?
A repetitive task. A disconnected process. An idea that needs a first version. Let’s start there.
Tell us about it