LEAVELEAVE
Get in touch

LEAVE AI & Data

Understand the scene.
Know the business.
Connect the next step.

From images and audio to location and business knowledge.
Bring vision, language models and data into everyday asset operations.

Explore our capabilities
Urban traffic and mobile assets
Built around real operationsPerception · Understanding · Coordination

Computer vision training

Turn a frame into
operational context.

Define targets, prepare annotated data, and train and evaluate models for vehicles, drivers and cargo. Connect recognition results with events and monitoring workflows.

A truck ahead on a mountain road, with vehicle and plate detection overlays
ADAS · Forward vehicle detectionExample detection
Vehicle 01 · TruckPlate JF·53·4542 m
License plate
JF·53·45
Distance ahead
42 m
Current speed
68 km/h
Time headway
2.2 s
Illustrative detection interface. Plate, distance, speed and headway are sample data.

Driving safety

Plan training and evaluation for distraction, fatigue, seat-belt use, phone use and smoking, according to available data.

Vehicles & cargo

Plan detection, classification and exception recognition for vehicles, plates, cargo and work areas.

From samples to deployment

Define sample and annotation standards, model evaluation and integration, with deployment suited to edge or cloud resources.

Discuss your vision use case

Agents & model fine-tuning

Bring business context
into every AI workflow.

Connect project knowledge, business data and tools for exception analysis, reports and operational assistance. Choose retrieval, fine-tuning and agent workflows to fit the task.

Choose a use case

Examples explain the service approach; no live data or model is connected.

Illustrative workflow

Why does this shipment need attention?

Connected context

Shipment record · Temperature events · Route history

Combine temperature excursions and stop records into a timeline, shipment milestones and items to investigate. Prepare a response for the operations team to review.

Output: exception summary, record references and suggested actions

Project knowledge retrieval

Organize manuals, procedures and business knowledge for retrieval with source references.

Domain fine-tuning

Prepare examples and evaluations for terminology, output formats and specific tasks; compare against a baseline before fine-tuning.

Agent development & integration

Connect authorized business interfaces and tools, with review steps for important actions and execution records.

Discuss agents and models

Data services

Bring every type of data
into your workflow.

Location, addresses, text, plates, faces and voiceprints. Connect devices, platforms and business systems through clearly defined inputs and outputs.

WiFi + LBS

Global assisted positioning

Use WiFi access-point and cellular base-station information to supplement mobile asset positioning worldwide when satellite visibility is limited.

  • Indoor and warehouse assistance
  • Location context at transport milestones
  • Location references for low-power assets

Integration & workflow

Collect available WiFi access-point and cellular observations at warehouses, indoor sites or loading areas, and use a request ID to link the shipment.

Store the estimated position alongside its source, quality information and capture time to supplement the journey history.

Example request

POST/v1/data/position/resolve
{
  "request_id": "DEMO-POS-1028",
  "wifi": [
    {
      "bssid": "02:00:00:00:00:01",
      "rssi": -62
    }
  ],
  "cell": [
    {
      "mcc": 460,
      "mnc": 0,
      "lac": 4101,
      "cell_id": 2073
    }
  ]
}

Illustrative contract; confirm paths and fields in the project API documentation.

Discuss integration
WiFi and cellular positioningInterface example

A cold-chain shipment enters a warehouse; WiFi and cellular data supplement its position.

Warehouse A · Cold-chain storageLoading baysWiFi AWiFi BLBSLBSCargo CC-1028
Business scenes and recognition results are illustrative.
Input4 WiFi access points · 2 cell towers
Example output31.2304, 121.4737
Positioning type
WiFi + LBS
Location result
Estimated position

Positioning and coverage depend on available access points, base stations and data conditions.

Deliver data where it is needed.

Connect Webhook, MQTT, SMS, email and Telegram / WhatsApp notifications with platform automation.

Explore data delivery

Data services

Review use cases, inputs and outputs to find the right service for your integration.

Global assisted positioning

Use WiFi access-point and cellular base-station information to supplement mobile asset positioning worldwide when satellite visibility is limited.

Input
Device-collected WiFi access points, cellular base stations and network information
Output
Coordinates, positioning type and available location-quality information

Global reverse geocoding

Turn latitude and longitude into readable geographic locations and addresses for asset history, alerts and reports.

Input
Longitude, latitude and requested language
Output
Geographic location, regional hierarchy and available address information

Form OCR

Extract text and key fields from transport forms, waybills and handover documents, with business validation to reduce manual entry.

Input
Form images or scans and field definitions
Output
Recognized text, structured fields and items requiring review

License plate recognition

Extract plate information from vehicle images or authorized video frames and link it to vehicle records, site access and shipments.

Input
Vehicle images or video snapshots
Output
Plate text, available recognition attributes and review information

Face recognition

Support identity verification and handover checks in authorized workflows using facial features and business rules.

Input
Authorized face images and comparison material
Output
Comparison result, reference information and human-review items

Voiceprint recognition

Extract speaker features from authorized recordings to support voice verification and related business records.

Input
Authorized audio samples and reference material
Output
Speaker comparison result and available quality information

Planning your solution

Practical guidance on device selection, coverage and integration before you start.

Discuss your project
How do I choose retrieval, fine-tuning or agents?

Retrieval supports answers grounded in documents. Fine-tuning needs task examples and baseline comparisons. Agents connect authorized tools and workflows. Define goals, data and acceptance criteria before selecting an approach.

Are the API examples production endpoints?

The JSON examples illustrate request structure and integration concepts. Confirm production endpoints, authorization, fields, coverage and acceptance criteria during project integration.

Collaboration & delivery

From defining the use case
to putting it to work.

Agree deliverables and acceptance criteria for each stage around your objectives, actual data and operating environment.

  1. 1

    Define the task

    Define targets, business actions, inputs and outputs.

  2. 2

    Prepare & validate

    Review real samples, data quality and a task baseline.

  3. 3

    Develop & integrate

    Integrate models, agents or data interfaces with the workflow.

  4. 4

    Review & refine

    Review against agreed criteria and refine with operational feedback.

Start with a specific need.

Discuss your project