Flitware

By FLITWARE S.A.S.

Overview

Flitware connects ChatGPT to your fleet operation: list vehicles and their status, review overdue preventive maintenance, inspect work orders, track document expirations, submit driver checklists, and open support tickets. Reads are tenant-isolated; writes use preview-then-confirm, and sensitive changes require human approval in the Flitware web app. Flitware only accesses data available to your connected account.

Tools

approval_status

ChatGPT

Poll a human approval proposal (tenant-scoped).

checklist_response_get

ChatGPT

Get one checklist response with answers snapshot, evaluation, observations and evidence refs.

checklist_response_list

ChatGPT

List tenant checklist responses with vehicle/driver/checklist/date/outcome filters.

checklist_stats

ChatGPT

Checklist scope counts (active items + groups).

checklist_submit

ChatGPT

Submit a driver checklist for a fleet vehicle.

contact_search

ChatGPT

Search tenant contacts for service creation (id + label only, isolated sub-query).

cost_compare_6m

ChatGPT

Six-month maintenance cost comparison (closed work orders, actual_cost).

document_expiring_list

ChatGPT

Documents expired or expiring within N days (metadata only).

document_get_download_url

ChatGPT

Issue a single-use download URL for one tenant document.

document_get_metadata

ChatGPT

Get one document metadata record (never file bytes or URLs).

document_list_metadata

ChatGPT

List document metadata (never file bytes or URLs).

document_related_list

ChatGPT

Documents related to a vehicle and/or driver, with validity at a date.

domain_catalogs

ChatGPT

Closed domain values: statuses, types, priorities, currencies and per-operation values.

driver_get

ChatGPT

Get one driver with assignment and masked contact.

driver_list

ChatGPT

List tenant drivers (operational fields, masked contact).

expense_duplicate_check

ChatGPT

Find duplicate expense candidates (read-only).

expense_register

ChatGPT

Register a fleet expense for a vehicle with duplicate check. Requires fleet expense accounts to be enabled on the tenant.

fleet_get

ChatGPT

Get one fleet vehicle by id (tenant-scoped).

fleet_list

ChatGPT

List fleet vehicles of the caller's tenant.

fleet_meter_history

ChatGPT

Odometer/hourmeter history of one fleet vehicle.

fleet_operational_summary

ChatGPT

Per-vehicle operational summary: documents, meter age, checklists, work orders, assignments. Sections gated by caller scopes.

fleet_readiness_check

ChatGPT

Vehicle readiness for a date: documents, maintenance due, checklists, conflicts, driver. PASS/FAIL/UNKNOWN per check.

job_cancel

ChatGPT

Cancel a PENDING/RETRY own job.

job_get

ChatGPT

Get one own job (tenant-scoped).

job_retry

ChatGPT

Retry a FAILED own job after reconcile.

maintenance_analysis

ChatGPT

Explained maintenance analytics: cost per km/hour, recurrent failures, vehicle comparison, MTTR by vendor, document risk, PM compliance detail.

maintenance_metrics

ChatGPT

Six-month maintenance dashboard aggregates.

maintenance_overview

ChatGPT

Operational maintenance counts.

maintenance_reliability

ChatGPT

Policy-backed reliability metrics (MTTR/MTBF/availability/PM compliance).

maintenance_schedule

ChatGPT

Preventive schedule states per vehicle.

maintenance_schedule_create

ChatGPT

Assign a resource to a work order (PLANNED, anti-overlap).

operations_review

ChatGPT

Daily operations review: blocks, warnings and info gaps with evidence and actions. Deterministic versioned rules.

operations_triage

ChatGPT

Triage daily operations findings into executable, pending approval, manual or blocked with evidence, deduped open work orders and next steps. Use when the operator asks what needs attention today or what to do about a finding. Example: {"horizon_days":7}.

plan_status

ChatGPT

Tenant subscription access mode (works in any mode).

plan_usage

ChatGPT

Tenant plan usage (ROOT only, full or recoverable).

portal_contact_read

ChatGPT

Read the portal contact's own profile and services.

portal_driver_read

ChatGPT

Read the portal driver's own profile, services and vehicle.

report_vehicle_costs

ChatGPT

Monthly cost report for one vehicle with chart hint, CSV and markdown fallback.

scheduled_run_get

ChatGPT

Get one scheduled execution with its delivery summary.

scheduled_run_list

ChatGPT

List execution history of a scheduled task (deliveries: summary + evidence + errors).

scheduled_task_create

ChatGPT

Create a persistent scheduled task (v1: read-only maintenance analyses).

scheduled_task_delete

ChatGPT

Delete a scheduled task and its execution history (owner or ROOT).

scheduled_task_get

ChatGPT

Get one scheduled task with its schedule, params and fire state.

scheduled_task_list

ChatGPT

List the tenant's scheduled MCP tasks (persistent automations, v1: read-only analyses).

scheduled_task_update

ChatGPT

Update a scheduled task (owner or ROOT): name, status (pause/resume), schedule or params.

search_global

ChatGPT

Cross-domain search with per-domain scope gate (zero hits and zero counts without scope). domains uses singular names (fleet, service, maintenance, document, checklist, support) and requires at least one.

server_discover

ChatGPT

Catalog of tools visible to the caller (scopes + plan + role).

service_candidate_evaluate

ChatGPT

Evaluate vehicle candidates for a service date: documents, maintenance, checklists, schedule clashes, driver and service fit, with 1-3 explained alternatives. Use when the operator must pick a vehicle for a service or a re-assignment. Example: {"date":"2026-10-30","fleet_ids":["<fleet_id>"]}.

service_create

ChatGPT

Create a PENDING service for a tenant contact.

service_get

ChatGPT

Get one service by id with PII-masked notes.

service_item_add

ChatGPT

Assign vehicle/driver/contact to a service via a service item (rich domain association; header holds a direct shortcut).

service_list

ChatGPT

List services of the tenant with PII-masked notes.

service_metrics

ChatGPT

Service aggregates for a month in the tenant timezone.

service_plan_apply

ChatGPT

Apply an approved service plan as one workflow (1..100 sealed steps): per-step revalidation just before apply, per-step and per-workflow idempotency, PARTIAL 207 with explicit manual compensations when some step fails. Use when the operator approved a plan from service_plan_compose and asks to assign the resources. Example preview: {"preview":true,"objective":"Organiza los servicios de manana","date":"2026-10-31"}. Example apply: {"sealed_preview":{...},"idempotency_key":"<opaque-key>"}. Preview first with objective (+ service_id/fleet_ids/date/driver_id); then apply the sealed preview with one idempotency_key (steps derive key+"#"+i); retry partials with the SAME key before snapshot_expires_at; never hand-edit sealed params.

service_plan_compose

ChatGPT

Compose an approvable service plan for a date: verified recommendation, PASS/FAIL/UNKNOWN, blockers, evidence, pending decisions and preview-first next actions, with a 15-minute snapshot expiry. Use when the operator asks to organize services or which vehicles can operate on a date. Example: {"objective":"Organiza los servicios de manana","date":"2026-10-30"}.

service_update

ChatGPT

Update a service: change dates, addresses, notes, price, or status. Validates allowed status transitions, requires a fresh read stamp, and needs human approval for price and status changes.