Text Tokens
Input / output tokens
Metered per model version, with prompt and completion counted separately.
Typical scenesEnquiry dialogue, document summarisation, declaration drafts, periodic reports
Multimodal Token Service
Usage Measured Per Modality, Reported As One Record
A single shipment file mixes documents, photographs, video clips and voice calls. Metering all of it as text tokens gives no usable view of where cost is generated. This service measures each modality on its own terms and converts everything into one comparable usage record.
Multimodal Token Service
Multimodal Token Metering & Billing
Industry workloads are rarely text-only: a single shipment file mixes documents, photographs, video clips and voice calls. This service meters each modality on its own terms and converts everything into one comparable usage record.
Text / image / video / voiceUnified conversion unitPer-modality quotasCross-modal task tracingModality-level audit log
Text Tokens
Dialogue, document summarisation and report generation metered by input and output tokens per model version.
Image Tokens
Document images, cargo condition photographs and yard stocktaking frames metered by resolution and processing depth.
Video Tokens
Handling surveillance, loading and unloading review and safety event clips metered by duration and sampling rate.
Voice Tokens
Multilingual service calls, transcription and voice broadcast metered by audio length and language pair.
Cross-Modal Task Metering
A task that reads a document, checks a photograph and drafts a reply is recorded as one traceable unit with its component costs.
Unified Conversion & Dashboard
All modalities converted to a comparable unit, with usage views by project, department, customer and modality.
Per-Modality Quota & Budget
Separate ceilings for image and video workloads, which are usually the fastest-growing cost line.
Caching & Down-Sampling
Repeated reference material served from cache, and imagery down-sampled where full resolution adds no decision value.
Modality-Level Audit
Call-level records retained per modality with model version and outcome, exportable for reconciliation.
Metering Basis
How Each Modality Is Counted
Counting rules are published rather than inferred from a bill, so a customer can estimate the cost of a workflow before it runs.
Image Tokens
Resolution & processing depth
Frames metered by resolution band and by whether the task is detection, classification or full field extraction.
Typical scenesDocument images, cargo condition photographs, yard stocktaking frames
Video Tokens
Duration & sampling rate
Clips metered by length and by how many frames per second are actually analysed.
Typical scenesHandling surveillance, loading and unloading review, safety event clips
Voice Tokens
Audio length & language pair
Calls and broadcasts metered by audio duration, with transcription and synthesis reported separately.
Typical scenesMultilingual service calls, transcription, voice broadcast and read-back
Governance
Control, Attribution and Audit Across Modalities
Unified Conversion Unit
Every modality is converted into one comparable unit, so a finance team reads a single usage line instead of reconciling four different measures.
Cross-Modal Task Tracing
A task that reads a document, checks a photograph and drafts a reply is recorded as one traceable unit, with component costs itemised underneath.
Per-Modality Quota & Budget
Separate ceilings for image and video workloads, which are usually the fastest-growing cost line and the easiest to overrun silently.
Caching & Down-Sampling
Repeated reference material is served from cache, and imagery is down-sampled where full resolution adds no decision value.
Modality-Level Audit Log
Call-level records retained per modality with model version, input reference and outcome, exportable for reconciliation.
Attribution by Project & Customer
Usage attributed to project, department and end customer, so recharge to the right cost centre does not require manual splitting.
Cross-Modal Workflows
Where Modalities Combine in One Task
Document Intake Desk
Image + TextA photographed invoice is parsed into structured fields, checked against the booking record and summarised for the officer. Image and text consumption are recorded under one task ID.
Yard Condition Review
Image + VideoStill frames from gate cameras are combined with short handling clips to evidence container condition, with sampling rate reduced for routine passes.
Multilingual Service Desk
Voice + TextAn inbound call is transcribed, answered from the knowledge base and confirmed by voice broadcast, with audio length and text tokens reported together.
Safety Event Handling
Video + TextA detected event produces a short clip, a structured record and a corrective action draft, metered as one cross-modal unit.
Talk To Us
Tell Us About Your Cargo
Share the origin, destination, cargo type and target schedule. Our team will review the details and reply with a proposed arrangement.
- AddressLumut, Perak, Malaysia
- TelephoneTelephone number to be confirmed
- EmailEmail address to be confirmed
- Working HoursMonday - Friday, 9:00 - 18:00 (MYT)