From Physical Robot Tasks to
Training-Ready Datasets

Mr. Robot AI Labs collects structured robot demonstration data using AgiBot G2 robots at our Data Collection Centre in Malaysia.

We support robotics and AI teams with task-specific data collection, annotation and dataset preparation through a managed workflow.

Talk to Our Team
View How It Works

Robot AI Needs
Real-World Task Data

Training robot models requires collecting high-quality real-world data, which can be expensive, slow and operationally complex to do in-house.

Mr. Robot AI Labs provides a managed solution for collecting and preparing robot task data, so your team can focus on model development, research and deployment planning.

From Task Collection to Dataset Delivery

We provide embodied AI training data through a managed DaaS model. Your team defines the task scope and data requirements.
Our team collects demonstrations, annotates the output and prepares the dataset based on the agreed structure.

Data collection with
AgiBot G2 robots

Real-world robot demonstrations collected through the AgiBot G2 wheeled humanoid robot platform.

Human-guided robot task demonstrations

Structured task demonstrations guided by trained operators in a controlled collection workflow.

Annotation through
Genie Studio

Collected data is annotated through Genie Studio based on agreed requirements.

Quality-checked dataset preparation

Datasets are reviewed before packaging and delivery.

Task scope aligned to client requirements

Collection scope is defined based on task type, volume, timeline and intended use.

Delivery Based on Agreed Requirements

Dataset delivery is prepared according to the agreed output structure and delivery method.

Data collection
with AgiBot G2

Real-world robot demonstrations collected through the AgiBot G2 wheeled humanoid robot platform.

Human-guided robot task demonstrations

Structured task demonstrations guided by trained operators in a controlled collection workflow.

Quality-checked dataset preparation

Datasets are reviewed before packaging and delivery.

Annotation through
Genie Studio

Collected data is annotated through Genie Studio based on agreed requirements.

Task scope aligned to client requirements

Collection scope is defined based on task type, volume, timeline and intended use.

Delivery to Specifications

Dataset delivery is prepared according to the agreed output structure and delivery method.

Designed for Robotics, Research and Automation Teams

DaaS is suitable for teams that need real-world robot demonstration data to support model training, research, testing or robotics development.

Robotic AI Companies

Academic and Research Institutions

AI Labs and Foundation Model Teams

Tech Companies Entering Humanoid Robotics

Automation and Manufacturing Teams

For teams developing manipulation, navigation or task-completion models that require structured robot demonstration datasets.

For institutions that need real-world robot demonstration data for robotics research, comparative studies or model evaluation.

For teams developing embodied AI models that rely on real-world task demonstrations for training, validation and performance improvement.

For companies exploring humanoid robotics and needing early-stage datasets to support testing, research or model development.

For industrial and commercial teams developing AI-driven robotic solutions for physical task execution, workflow automation or applied robotics use cases.

Real-World Task Data for Robot Training

The service is designed for robot demonstration data. Final scope depends on task complexity, required volume, output format and collection setup.

Physical Task Demos

Object handling, pick-and-place, sorting and other physical task demonstrations.

Navigation Sequences

Movement and path-following data for robot navigation and spatial understanding.

Interaction Sequences

Structured robot-object and robot-environment interaction demonstrations.

Custom Task Data

Task-specific collection scoped around project requirements and intended use.

Physical Task Demos

Object handling, pick-and-place, sorting and other physical task demonstrations.

Navigation Sequences

Movement and path-following data for robot navigation and spatial understanding.

Interaction Sequences

Structured robot-object and robot-environment interaction demonstrations.

Custom Task Data

Task-specific collection scoped around project requirements and intended use.

Dedicated Data Collection Centre

Mr. Robot AI Labs operates a Data Collection Centre in Malaysia to support structured robot demonstration data collection. The facility enables task demonstrations to be carried out in a controlled environment with trained operators and a defined collection workflow.

The service uses AgiBot G2 wheeled humanoid robots and Genie Studio for annotation.

Applications for Robotics AI Development

The collected data can support robotics AI teams across model training, research, benchmarking and applied development.

Manipulation Model Training

Demonstration data for tasks such as grasping, sorting, placing and object handling.

Navigation and Spatial Awareness

Movement & path tracking for robot navigation, positioning & spatial task understanding.

Task-Completion Demonstrations

Structured demonstrations that show how physical tasks are performed start to finish.

Benchmarking Datasets

Standardized datasets that help teams compare model performance across defined tasks.

Generalized Robot Task Research

Research data for broader robot behaviour, task policies and embodied AI development.

Industrial and Commercial Robotics Development

Task-specific data for robotics uses in industrial and/or commercial environments.

Mr Robot AI Labs

DaaS End-To-End Workflow

1. Submit Your Data Request

Share the task type, expected volume and preferred output requirements.

2. Confirm Scope

Our team aligns the task, collection setup, timeline and delivery requirements.

3. Configure the Collection Environment

The DCC setup is prepared based on the agreed tasks.

4. Run Robot Demonstrations

Robot Trainers operate AgiBot G2 robots through the defined tasks.

5. Capture and Process Data

Demonstration data is captured and prepared for annotation.

6. Annotate Through Genie Studio

The collected data is labelled based on the agreed requirements.

7. Quality Check and Review

The dataset is reviewed before packaging.

8. Deliver the Structured Dataset

Your team receives the agreed dataset for model training, research or benchmarking.

Talk to Our Team

Scoping Requirements for Data Collection

Every project is scoped before collection starts. This helps both teams align on the task, data requirements, structure and delivery expectations.

Every project is scoped before collection starts. This helps both teams align on the task,
data requirements, output structure and delivery expectations.

Task Requirements

Task type, task complexity, collection environment and expected demonstration scope.

Data Requirements

Data volume, data type, annotation requirements and preferred output structure.

Delivery Requirements

Output format, delivery method, timeline and review process.

Usage Scope

Usage rights, commercial scope, project timeline and any special requirements.

Technical Follow-Up Note

Specific details such as data formats, delivery structure, sensor specifications, API availability and usage terms can be confirmed during the scope discussion.

View the DaaS Deck

Prefer a slide-format overview? View the Mr. Robot AI Labs DaaS deck
for a broader summary of the service, data products and collection model.

View the Mr. Robot AI Labs DaaS Deck View DaaS Deck

Frequently Asked Questions

What is DaaS?

DaaS means Data as a Service. For this service, it refers to the collection, annotation and preparation of robot task data based on agreed project requirements.

Who is this service for?

This service is suitable for robotics companies, AI labs, research institutions and technology teams that require real-world robot task data for training, testing or research.

What robot platform is used?

The service currently uses AgiBot G2 wheeled humanoid robots at Mr. Robot AI Labs’ Data Collection Centre in Malaysia.

Can we request custom task data?

Yes. Custom task data can be discussed based on the task type, required volume, collection setup and delivery requirements.

What data formats are available?

Output formats are confirmed during scoping. You may share your preferred format in the enquiry form so our team can review the requirement properly.

How is the data annotated?

Collected data is annotated through Genie Studio based on the agreed project requirements.

Is API delivery available?

Delivery methods are confirmed during scoping. If your team requires API or system integration support, please include this in the enquiry.

How are data rights handled?

Data rights and usage terms are confirmed before collection begins.

Talk to Our Team

Share your question, requirement or project interest with Mr. Robot AI Labs. Our team will review your enquiry and follow up accordingly.