Caliber is the field-grade scoring platform used by selection cadres to capture, structure, and compare candidate performance across physical, cognitive, and sociometric tasks, with the rigor that elite-unit decisions demand.
Caliber was built to replace the clipboards, spreadsheets, and improvised group chats that selection cadres still rely on during multi-day assessment events. It was designed in close cooperation with active instructors and supervisors, around the way evaluation actually happens on the ground, not the way a desk-built tool assumes it does.
The platform structures an entire field selection process from the moment candidates form a group through final ranking: timed sprints with full kit, navigation drills, obstacle courses, cognitive tasks, peer assessments, and instructor notes, all captured against the same candidate record, on the same device, by the people who saw the performance firsthand.
Caliber is operated by instructors, not analysts. Every interaction is tuned for one-handed use under load, in glare, in dust, often without connectivity. Data is held locally on the device, encrypted at rest, and synchronized when the network allows, never the other way around.
The system supports multiple organizations and units from a single deployment, with role-based access for instructors, supervisors, and event commanders. Each event runs against its own configured battery of tasks, scoring rubrics, and group composition rules, so the same platform serves different selection doctrines without forcing them to converge.
Caliber is opinionated about what evaluation requires and stays out of the way of what it doesn't. Every feature exists because an instructor asked for it on a real event.
Compose groups, manage rosters, attach notes and metadata to every candidate. Re-organize on the fly when the event demands it.
25+ predefined game types covering sprints with kit, navigation, obstacle work, and cognitive tasks, plus user-defined custom tasks with their own rubrics.
Free-text notes attached to any candidate, any moment. Searchable, time-stamped, exportable, built for the de-brief, not just the score.
Written for the field. Every action, including capture, score, and notes, works fully offline. Synchronization is a background concern, never a precondition.
Instructors evaluating the same group can share results directly between devices, even with no network. Scores merge into a single record without funneling through a server.
Dictate observations hands-free, or tap a predefined comment from the cadre's standardized library. The right note ends up on the right candidate without slowing the task down.
Every candidate ends the event with a structured summary: scores per classification, narrative description, and a single final grade. The whole picture, on one screen, ready to defend in the de-brief.
A live web dashboard tracks every group submission across the event. Pass, limit, and fail counts are aggregated in real time, while score gaps between instructors evaluating the same candidate are flagged automatically for review.
One deployment, many tenants. Each organization defines its own events, tasks, and scoring conventions. Data is strictly partitioned by org and role.
Selection data is sensitive. Caliber is engineered around that fact, with defense in depth from the device to the cloud, and cryptographic guarantees rather than policy ones.
Backups never leave the device unencrypted. Restoration requires the original PIN and the matching event metadata, knowing one without the other gets you nothing. Phone-based authentication is enforced for the Pro deployment, and every API call carries a verified Firebase ID token.
Caliber tracks the entire lifecycle of a selection event, and gets out of the way once you're running it.
The supervisor defines the event: tasks in play, scoring rubric, passing threshold, and how groups should be composed. Saved once, used everywhere.
Instructors compose groups on the device, capture candidate identities, and lock the roster. The same screen tracks attendance throughout the event: who's active, who's at the clinic, who withdrew.
Instructors lead each candidate through the battery of timed and scored tasks. The clock runs on the device, attempts are logged, and the next task is one tap away.
Per-candidate scoring on every task, plus instructor notes captured in real time by voice or quick-comment. Each candidate ends the event with a structured assessment summary.
The web admin panel tracks every group submission live: total pass, limit, and fail counts across the event, group-level approval workflow, and automatic flags on score gaps between two instructors evaluating the same candidate.
