Skip to content
Human knowledge. Machine possibility. Explore the robotics lab
9jabots.Create a task pack
HUMAN TASK DATA FOR AI & ROBOTS

Nigerian know-how.
AI-ready context.

Turn everyday Nigerian tasks into structured evidence for AI and robotics.

Annotate your videos, preserve English and Pidgin instructions, and export task data your team can inspect and test. Start in your browser.

Try the task studio
TASK STUDIO EARLY RELEASE / 0.1
Draft saved on this device
THE DEMONSTRATIONAUTHORED TEMPLATE
WATERWATEROBJECT_01 · SOFT PACKAGETARGET · SERVING TRAYTASK STUDY / ILLUSTRATIONHUMAN ACTION → STRUCTURED KNOWLEDGE
Example task plan · no recorded footage

Small details are the data. Soft packaging, familiar phrases, tight spaces. Capture the context an agent needs to get the task right.

MAKE THE ACTION EXPLICIT

Serve sachet water

STEP 01 OF 4
Structure checks pass
WORKS IN YOUR BROWSER

Videos stay on your device. Annotation drafts are saved here. Your data controls.

GROUNDED IN A WORKING ECOSYSTEM9jaTesters Human insightNGPT Local languages9jaBench Measured evidence9jaRobotics Embodied systems
01 / START WITH SOMETHING YOU CAN INSPECT

Local language.
Testable instructions.

Can a system understand where to go,
when to stop, and when to ask?

INDOOR LANGUAGE / V0.1

Instructions with a place in the world.

24 authored examples of Nigerian English and Pidgin instructions, grounded in a named indoor environment. Includes corrections, stops and ambiguous requests.

24 task fixtures2 language labelsJSON + dataset card

Authored starter fixtures. Explore instructions, expected actions and context before designing your own collection.

Download the sampleRun the benchmark
NIGERIA MOBILITY / COLLECTION WORKFLOW

Plan the next 100 observations.

A 100-task assignment pack for road scenes, local objects, charger observations and quality review. Use it to plan a documented collection with your team.

02 / A LOCAL ADVANTAGE, MADE REUSABLE

The details change
what a machine must learn.

A soft water sachet is not a rigid bottle.
A familiar phrase can carry an unfamiliar instruction.

01 / SHOW THE WORK

Human demonstrations.

Record actual tasks, objects and environments. The 9jaTesters network can contribute and review local data under documented collection terms.

02 / MAKE IT LEGIBLE

Context, not just clips.

Attach actions, objects, language, time ranges and expected outcomes. Preserve provenance so another team knows what the data can support.

03 / TEST WHAT TRANSFERS

Evidence for AI and robots.

Evaluate task understanding and navigation first. Add synchronized sensor and robot-action data when hardware collection is available.

03 / FIELD NOTES

Build with the context.

All field notes
DATA COLLECTION

What makes Nigerian task data useful for robot learning?

PRACTICAL GUIDE

From a phone video to a documented task pack.

LANGUAGE & CONTEXT

Test whether an AI understood the instruction.

BUILT BY RANKED TECHNOLOGIES LTD

A working foundation.
A longer horizon.

Founded by author and builder Eruo Fredoline. Ranked.ng and 9jaTesters already serve paying customers. 9jaBots extends that foundation into human task knowledge for the next generation of AI and robotics.

Explore Ranked Meet the data network
FROM A HUMAN TASK TO A MACHINE TEST

Now, put it to the test.

Inspect model predictions. Change a simulated environment. Find the failure worth fixing.

Enter 9jaRobotics