Inventory, logistics, and operations

Turn operational images into counts your team can review

Count repeated objects, pallets, cartons, and materials from clear photographs, inspect the marked result, and integrate structured counts where appropriate.

  • Web products available now
  • Batch only where labelled
  • API only where documented
Organised warehouse shelves with pallets, cartons, and repeated products ready for visual counting
Workflow illustration

Input

Source files

Output

Workflow-ready result

An illustrative view of the inventory operations workflow.

Industry challenges

The operational problem starts before the first edit

Visual counting works best when the capture process is designed for counting. Occlusion, perspective, mixed items, and poor lighting create exceptions that a result screen must make reviewable.

01

Manual tallies are repetitive

Teams repeatedly count visible stock, packages, bundles, or materials from shelves and staging areas.

02

The image controls the result

Crowded scenes, hidden objects, glare, and mixed scales can make any visual count uncertain.

03

A number without evidence is hard to trust

Operations teams need marked items or returned coordinates so they can review what the system counted.

End-to-end workflow

A reviewable image-count workflow

Capture quality, target selection, counting, and exception review are all part of the workflow—not just the final number.

  1. 1

    Define the count target

    Decide whether the image should count all repeated objects, people, or a specific item class.

  2. 2

    Capture a countable scene

    Use clear light, visible separation, and a useful angle; split heavily occluded areas into more than one image.

  3. 3

    Choose the relevant counter

    Use the general object counter or a product-oriented workflow such as inventory, pallets, boxes, or rebar.

    Object Counter
  4. 4

    Run the count

    Process a single image or use registered web batch modes for repeated operational images.

    AI Inventory Counter
  5. 5

    Review the marked result

    Compare visible objects and returned markers before the count is used in an operational record.

  6. 6

    Export or integrate

    Use the web result or send a controlled image to the public counting API for a structured response.

Relevant products

Counters selected for operational scenes

Specialist pages guide the capture and review experience; the public API exposes object and people targets rather than a separate endpoint for every item.

Pixoate product

AI Inventory Counter

Frames the workflow around repeatable stock-taking images and reviewable outputs.

  • Web app
  • Web batch
  • Public API
Count Inventory

Pixoate product

Object Counter

Provides the general still-image counting workflow and maps directly to the public object target.

  • Web app
  • Web batch
  • Public API
Count Objects

Pixoate product

Pallet Counter

Gives warehouse teams a focused entry point for visible palletized units.

  • Web app
  • Web batch
  • Public API
Count Pallets

Pixoate product

Box & Carton Counter

Supports visible carton and package-count scenarios from clear operational images.

  • Web app
  • Web batch
  • Public API
Count Boxes

Pixoate product

Rebar & Rod Counter

Provides a focused workflow for repeated material ends in construction and stock scenes.

  • Web app
  • Web batch
  • Public API
Count Rebar

Counting preview

The result includes evidence, not just a total

Review item markers in the web app or use returned normalized centers from the API to inspect what contributed to the count.

Operational image

One defined target with visible, separated items

Reviewable count

Count, engine, item centers, and optional annotated output

API and automation

Return structured object counts to an inventory system

The public counting endpoint accepts a still image and returns the count plus normalized item centers. Optional parameters control target, mode, sensitivity, exemplars, and annotated output.

Input
A multipart image, object/people target, optional mode and controls, API key, and idempotency key.
Processing
Pixoate selects or runs the requested counting mode as a synchronous metered request.
Output
JSON with target, count, engine, normalized points, credits spent, and optional annotated-image references.
/api/v1/countAvailable now

Sample request

POST /api/v1/count
X-API-Key: <your-key>
Idempotency-Key: <unique-key>

image=@shelf.jpg
target=objects
mode=auto
includeImage=1

Sample response

{"success":true,"target":"objects","count":94,"engine":"<engine>","points":[[0.512,0.203]],"creditsSpent":1}

Current boundary: The endpoint is synchronous and image-based. It does not provide a warehouse-management integration, autonomous stock reconciliation, public async jobs, or webhooks.

Workflow value

Counting workflow measures to track

Start from your current baseline, run the workflow, and measure these operational indicators in your own environment.

Scenes counted per review cycle
Use consistent capture rules and registered web batches for repeated images.
Counts requiring manual correction
Inspect marked results and improve capture conditions for recurring exception types.
Time spent on first-pass tallies
Begin review from a machine-generated count instead of a blank tally sheet.
Evidence retained with a count
Pair the operational number with visible markers or returned point coordinates.

Product evidence

Capabilities you can validate directly in Pixoate

Focused web products

General and item-specific counting routes exist for objects, people, inventory, pallets, boxes, rebar, and other maintained verticals.

Batch counting is registered

The maintained batch registry includes the object counter and multiple operational counting products.

Structured public counting response

`/api/v1/count` returns a count, engine, and normalized centers, with optional annotated-image output.

Practical answers

Inventory and object-counting questions

What makes an image suitable for inventory counting?

Items should be visible, reasonably separated, and captured in usable light. Split heavily occluded or very large scenes into multiple controlled images and review every result.

Can Pixoate count a specific repeated item?

The object API supports optional exemplar boxes for similar-item counting, and the web app provides focused item workflows. Results still depend on the source scene.

Does the API return item locations?

Yes. The documented count response includes normalized center points for counted items and can optionally return an annotated image reference.

Can Pixoate update our warehouse system automatically?

Pixoate returns the structured count. Your system must decide how to validate, map, and apply it; no named WMS integration or autonomous reconciliation is advertised.

Is bulk counting available?

Relevant web counters are registered for Premium batch workflows. The current public API does not expose an asynchronous multi-image job endpoint.

Product limits and access can change by route and plan. The linked product, pricing, and developer pages remain the final source.

Related solutions

Follow the workflow into adjacent teams

Explore all solutions
Receipts and finance documents flowing into organized spreadsheet-style data
Finance & Accounting

Finance document automation

Move receipts, tables, statements, and scanned documents into reviewable text and spreadsheet workflows.

Explore solution
Varied seller product photographs flowing into a consistent marketplace catalog grid
Online Marketplaces

Marketplace image automation

Standardise seller-uploaded imagery without turning every listing into a manual design task.

Explore solution
Images flowing through a polished processing core into clean visual assets and structured results
Technology & SaaS

Image processing APIs for SaaS

Embed selected image transformations, extraction, and counting in a product you already operate.

Explore solution

Choose your path

Make the first count faster—and the review clearer

Try a focused web counter, plan a repeatable capture process for your team, or integrate the structured object-counting response.