Bookings, lifecycle, scheduling, and trip health.
Pre-flight checks that block a driver from marking a stop as arrived / finished when real-world conditions are not met —
Models the cargo itself — a Package with weight/volume/dimensions, tracking codes, plain recipient contact fields (no Cu
Batch-plans the day's pending PACKAGE segments into sequenced van-day routes respecting time windows and weight/volume/d
A logistics/delivery operating mode alongside taxi/people-movement, built by extending the existing Trip spine rather th
Baseline-driven health resolution engine (ported from the decommissioned go-agent)
Customer-fault miles penalty when a customer cancels late: a no-show penalty if the driver had already arrived, otherwis
A single Trip can carry multiple chained segments (multi-load), enabling shared-ride, ride-pool, and school-run patterns
Compliance chain a driver must satisfy before a stop can be marked no-show: must be near the pickup, must have waited th
Single lightweight endpoint that returns badge counts for every notification source the operator console renders — healt
A planned extension that moves the optimiser from combining trips to assigning them — allocating trips to specific drive
A planned extension that lets the optimiser react to in-flight events — cancellations, no-shows, delays — by replanning
Groups eligible trips into evaluation pools before they're optimised, so the engine only ever compares genuinely combina
Hierarchical merge-preference resolution: each actor (Customer, Account, AccountDepartment, Trip) can declare a MergePre
A planned extension that reshuffles work at the segment level: instead of merging whole trips, it can take segments from
Captures proof of completion — signature, photo, delivery code, plus who/when — against the Package at the delivery stop
Real-world routing for any two points — driving distance, duration and turn-by-turn geometry
A TripSchedule links a master Trip to a recurrence pattern (TripSchedulePlan) so the platform clones the trip on every m
Append-only record of every release-timing decision for a trip — the initial CREATE/UPDATE from the learned release-time
Operator-managed groups that pause one or many recurring trips for known disruptions — bad weather, voting days, school
Extra safety monitoring for a special-category passenger once they are on board — tighter alerts for a driver going GPS-
Driver-initiated counterpart to automatic zone entry: a driver declares the zone they are driving to and reserves a queu
Per-stop time windows constrain when a pickup or dropoff may happen — a customer's "between 9 and 10am at the school", a
Heatmap endpoint returning per-H3-cell outbound traffic congestion factors at a given day-type + time-slot
Guarantees a shared ride never exceeds the vehicle's capacity
Time-bound soft indicator that an operator is currently editing a trip
Architectural overview of the lifecycle-guard subsystem: pre-flight checks that block or surface advice on a trip transi
HTTP API powering the operator console health dashboard: traffic-light summary of WARNING / CRITICAL counts across all a
Continuous monitoring of every active trip for risk conditions — late prediction, stale GPS, allocation risk, tight wind
Architectural overview of the trip-health subsystem: how checks are evaluated each sweep, the WARNING / CRITICAL severit
Reusable, at-any-stage answer to "is this trip running late, and by how much" plus a SYSTEM/DRIVER/CUSTOMER fault decomp
An end-to-end view of how a trip moves through its states — from booking through allocation, dispatch and in-progress to
The continuous optimisation sweep
An overview of how the platform combines and sequences trips for efficiency: the merge-strategy hierarchy that decides w
HTTP endpoints to manually trigger booking-time preauth, trip-end capture, or refund on a specific segment
How a booking becomes the concrete work a driver carries out — resolving it into executable units, deciding how segments
A unit of work is one self-contained run of driver work — typically an uninterrupted sequence of stops
Operator-defined categories that classify trips based on a JSON decision tree (e.g
Records every booking that was rejected before becoming a Trip — explicit operator decline, no-supply (no driver availab
How the platform reasons about time across a trip — the derived timing metrics (such as acceptable arrival and dispatch
A TripSegment is one leg of a Trip — the work between two stops with its own pickup, dropoff, capacity requirements, lif
A TripSegmentStop is one waypoint on a segment — pickup, dropoff, or intermediate stop
A comprehensive read-only view of a single trip — timeline, driver/vehicle, account + custom fields, per-segment stops w
A Trip is the top-level booking record on the platform
Automatically places drivers into the zones whose areas they enter