에이아이웍스 AIWORKX

모바일메뉴열기
blackolive

From autonomous driving to vision AI,
A new standard for data labeling

From images to massive LiDAR point clouds, experience precise labeling and efficient data management on a single platform.

  • Patent pending for large-scale 3DFiles with 30M+ points
    rendered in high quality in the browser
  • Data Voucher supplierAdoptable through government funding
  • Defense · Public air-gapped deploymentsInstalled and run on isolated networks
  • Data pipeline automationFrom collection to labeling, review, and delivery,
    connected automatically as a workflow
Trusted by

Leading Korean enterprises and institutions choose blackolive

  • Defense Agency for Technology and Quality
  • Hyundai Rotem
  • ROK Army AI Center
  • Korea Testing Laboratory
  • Hyundai Mobis
  • Samsung C&T
  • ROK Navy Data Center

11yrs

Focused on data and quality

Expertise built since our founding in 2015

6yrs running

Data Voucher supplier

Selected for the K-DATA Data Voucher program

3yrs running

AI+X Top 100

Named to the AIIA Emerging AI+X Top 100

Why blackolive

Why teams choose blackolive

Korea’s first to support 3D segmentation labeling
at the scale of tens of millions of points
AI labels first,
so annotators can focus on review and correction
3D labels are projected onto camera images to
review 2D and 3D in a single pass
From log extraction to final datasets,
automate the entire pipeline.

Labeling Tools

From 2D to 3D,
every labeling tool you need

  • Image tagging · Attribute classification
  • Hierarchical object labeling
  • Object tracking · Keyframe interpolation
  • Batch labeling
  • Personal data de-identification
  • Role-based access control
  • Multilingual interface
  • Project backup · Restore
  • External system API integration
  • Custom output formats
Project Management

From data labeling to project operations,
all connected on one platform

Project management, annotator assignment, quality review, and operations analytics. blackolive brings together everything large-scale AI data projects need for efficient, well-organized operations.

blackolive
blackolive project list screen
blackolive
blackolive task list in project details
blackolive
blackolive 3D LiDAR labeling screen
blackolive
blackolive project status screen
  • Integrated project management

    Manage even complex data projects efficiently on one platform.

    Manage each project’s data and task status in one place, and run everything from annotator assignment to schedules and progress in an organized way. Auto-assignment and ETA tracking raise operational efficiency for large-scale projects.

  • An efficient labeling workspace

    A workspace built for focus raises data quality.

    A labeling environment optimized for every data type, from 2D images to 3D point clouds. Check task progress and review status at a glance, and keep work continuous with draft saving.

  • Structured quality control and validation

    A structured review process that raises data quality.

    Multi-stage review and approval keeps labeling quality under control. Built-in checks such as GT-based accuracy scoring and BBOX distribution analysis, with results delivered as reports, make your data more reliable.

    Issue trackingAI auto-reviewGT comparison
  • Data-driven operations analytics

    See project performance and productivity in data.

    Progress by project and annotator, time spent per task, and statistics by labeling type are visualized for you. Key operating metrics and automation usage are analyzed to support efficient operations and decision-making.

  • Large-scale task auto-creation

    Spend less time preparing repetitive tasks and build data faster.

    Bulk data upload and template-based task creation let you generate large labeling batches at once, cutting the time and effort needed to set up a project.

  • Annotator training and evaluation with Academy projects

    Train in the same environment as real work, with results scored automatically.

    When annotators complete training assignments in an Academy project, their work is compared with GT and automatically scored for accuracy, precision, and recall. The results confirm each annotator’s expertise and keep labeling quality consistent.

Applications

The training data
each industry needs

Automotive · Automation Workflow

From driving data to AI training data,
delivered through an automated pipeline

From data preprocessing to labeling task creation, quality validation, and final dataset delivery. Each step is linked to fit your environment, making data production more efficient and consistent.

Auto-create labeling tasks from driving data

Select a driving segment and the labeling task is registered automatically.

  • Automatic labeling task creationSegment selection, clip creation, and task registration linked automatically
  • Shorter task preparationMinimizes repetitive data preparation
  • Check sensor alignmentOverlay LiDAR points on video to verify camera alignment
Task list screen: Cuboid output information and Cuboid inspection results (counts and object/frame lists for missing, late start/early end, not inspected, and occluded/out of view)
Automated labeling quality validation

Cuboid labeling results are inspected automatically.

  • Automatic error detectionChecks for missing objects, start/end timing errors, and occlusion/out of view
  • Find fixes fastSee items that need correction by type
  • Consistent quality controlLess manual review, one standard for validation
Data status screen: per-stage totals for clip creation, Cuboid labeling, SEG labeling, and final data, with progress by clip
Data processing status at a glance

Manage the entire data production process on one screen.

  • Track overall progressSee each item’s processing stage and status at a glance
  • Run follow-up tasks in bulkSelect multiple tasks and run the next step at once
  • Prevent missed tasksMake sure no unprocessed data slips through
End-to-end flowFrom data prep to final output, in three steps
STEP 01Data preprocessing and labeling task creation

Automatically handles everything from splitting driving logs and extracting target segments to registering LiDAR data and creating Cuboid labeling tasks.

  • Auto-split driving logs into 1-minute units
  • Select target segments and create clips
  • Auto-register LiDAR data
  • Create Cuboid labeling tasks
blackoliveCuboid labeling
STEP 02Labeling quality validation and process linking

Collects Cuboid labeling results automatically, runs frame-by-frame quality checks, and hands off to segmentation labeling tasks.

  • Auto-collect Cuboid labeling results
  • Auto-check missing, start/end errors, and occlusion
  • Convert data for segmentation
  • Create segmentation labeling tasks
blackoliveSegmentation labeling
STEP 03AI training dataset creation and validation

Collects segmentation labeling results, converts them into training-ready datasets, and supports final validation.

  • Auto-collect segmentation results
  • Convert to AI training datasets
  • Visualize and validate final outputs
Talk to us about a pipeline
aiworks plus

Technology meets expert talent,
raising the bar for data quality

Integrated with aiworks plus, we run the workforce behind your data,
from recruiting and training expert labelers to project assignment and quality review.

Crowdsourced workforce recruiting

We recruit data labelers suited to your project on the aiworks plus platform.

Academy expert training

Online and offline labeler training builds the skills your project requires.

Linked to blackolive projects

Trained specialists process your data in blackolive.

Expert review and quality control

A review process grounded in software QA experience keeps data quality high.

Customer Stories

Customers across industries
choose blackolive

High-quality training data is essential to ensuring the safety of the fully autonomous driving we are researching. Working with AIWORKX, we secured high-quality autonomous driving training data through the meticulous processing and review of specialists using the blackolive data processing tool. We could feel the experience and know-how AIWORKX has built over many years of producing autonomous driving data.
Seoul National UniversitySeung-Woo Seo
Processing harmful content for content filtering is hard, because it means continuously viewing and handling that content. AIWORKX successfully processed our harmful content data, backed by deep data processing expertise and a strong sense of responsibility. With AIWORKX, we laid the groundwork for healthier digital advertising.
ADOPWon-seop Lee, CEO
To simplify competency assessment with AI, we needed to process data so AI could learn from it, which required a great deal of time and skilled staff. AIWORKX’s text data processing solution helped us build exactly the data we were aiming for. Joining the Data Voucher program with AIWORKX was an excellent choice.
Leaders InsightDae-young Lim, CEO
Building an AI that manages your diet from a single photo starts with securing high-quality diet data, which means entering nutrition information for countless food images. Through the Data Voucher program with AIWORKX, we efficiently built the high-quality data we needed. It was great working with the AIWORKX PM, who showed professional project management built on constant communication.
SprintJong-in Kim, CEO
Deployment

From SaaS to on-premises,
deployment options that fit your environment

Recommended SaaS

Cloud subscription

Pricing
Monthly subscription (by seats · storage)
Data location
blackolive cloud
Updates
New features applied automatically
Best for
PoCs, research institutes, Data Voucher projects
Contact sales
Recommended PRIVATE CLOUD

Private cloud

Pricing
Annual contract
Data location
Your dedicated cloud (incl. public-sector cloud)
Updates
Deployed on an agreed schedule
Best for
Enterprises, public-sector cloud projects
Request a quote
Recommended ON-PREMISE

On-premises · Air-gapped

Pricing
License + annual maintenance
Data location
Your in-house servers
Updates
Offline installation packages provided
Best for
Defense, public, and finance on isolated networks
Discuss deployment

Pricing and terms vary with data volume and contract length. Contact us for a detailed quote.

Not sure which deployment fits?

FAQ

Frequently asked questions

QWhat types of data can be labeled?

Everything from images to massive LiDAR point clouds can be labeled on one platform. With 2D and 3D labeling tools for industry projects such as autonomous driving, defense and aerospace, and LLMs, you never need a separate tool when the data type changes.

QCan it handle large-scale 2D and 3D data efficiently?

Even 3D data with tens of millions of points can be worked on directly in a web browser, with no extra software to install. Create large batches of labeling tasks at once and manage progress in one place, so operations stay manageable as projects grow.

QCan it run on in-house servers or air-gapped networks?

Yes. You can deploy it in the way that fits your security policy, from cloud subscription to private cloud to in-house or air-gapped installation. The same features are available even where data cannot leave your network.

QCan it be automated around our existing data production process?

Yes. We build automated pipelines around your workflow, from data preprocessing and labeling task creation to quality validation and final dataset generation. Fewer repetitive tasks mean shorter production timelines and consistent results.

QHow is labeling quality managed?

Labeling and review are separate stages, and only data that a reviewer has checked and approved moves to the next step. Automated quality checks catch errors people easily miss, reducing review effort while keeping data quality consistent.

QHow are cost and deployment time determined?

They depend on the deployment option, data volume, and number of users. Share your project requirements and our team will review them and recommend the best deployment option, quote, and schedule. Adoption through the Data Voucher program is also available.

Contact

Contact us

Share your project requirements and our team will get back to you with deployment options, timeline, and pricing.

Purpose: customer consultation, handling complaints and requests, newsletters, and promotional information such as marketing and promotions · Retention: 1 year from the date of consent
For other details, please see our Privacy Policy. You may refuse consent to the collection and use of personal information, but doing so may limit your use of this service.

(주)에이아이웍스 대표이사 : 윤석원 전화 02-423-5178 Fax 02-424-5178

본사 주소 : 경기도 화성시 삼성1로5길 6, 813호, 814호. 18449

서울 연구소 : 서울시 송파구 백제고분로 41길, 42-19, 2F~4F. 05623

© 2025 AIWORKX. All rights reserved. 문의 cs@aiworkx.ai