INTELLECTUAL ASSET / 01T
Taste & judgement
What good means for a particular medium, audience, brand and narrative purpose. Expert craft, project feedback, preference and accepted outcomes keep the definition grounded.
ONE2X DATA / FOR AI NATIVE CONTENT
CONTENT DATAFactory.
We know content. We scale what makes it good.
One2X builds AI Native content products. Separately, our data business runs an expert production environment to create datasets, evaluations and training signals from project-approved or One2X-owned inputs—with taste, lineage and repeatable quality control.
DATA SERVICE / DATASET · EVALUATION · TRAINING SIGNALS · API
CONTENT DATA FACTORYALWAYS ON
CONTENT LOOP × TASTE LOOP → TRAINING DATA
A DATA SERVICE BUILT INSIDE A WORKING CONTENT ENVIRONMENT.
01CONTENT
02JUDGEMENT
03RECIPES
04SIGNALS
Content expertise
makes better data.
These are not showreel decorations. Each production case helps us define a task, isolate control variables, set a quality bar and expand one good result into paired data, variant groups, trajectories and preferences.
SELECTED CONTENT ENVIRONMENT OUTPUTS / OWNED OR PROVIDED FOR ONE2X USECONTROL · LINEAGE · JUDGEMENT · SCALE
02 / TWO FLYWHEELS
Content and taste generate the data.
The content loop gives us high-value tasks, finished media and real outcomes. The taste loop gives us judgement, critique, revision and preference. Running both inside the same environment turns content production into a scalable data factory.
One2X Data packages what the loops produce as datasets, evaluation services, training signals and APIs—without pretending that raw media alone is the product.
THE DIFFERENCE
CONTENT ENVIRONMENT × TASTE SYSTEM × DATA ENGINEERING
CONTENT ENVIRONMENT × TASTE SYSTEM × DATA ENGINEERING
TWO LOOPS / ONE FACTORYCONTENT-NATIVE DATA SERVICE
DATASETMedia, pairs, variants, trajectories and labels
EVALUATIONTask, rubric, expert judgement and outcomes
TRAINING SIGNALSCritique, correction, revision and preference
APIRepeatable tasks and continuously refreshed signals
03 / THE THINKING ASSET
Taste is not a moodboard. It is a system of decisions.
The most valuable production data is often invisible: why a shot works, where pacing breaks, which revision earns approval, and what an audience responds to. One2X Data makes those decisions explicit, structured and deliverable.
INTELLECTUAL ASSET / 02G
Narrative & media grammar
Shot logic, pacing, continuity, visual hierarchy, article–image relationships and layout. The rules change by medium; the system keeps them explicit.
INTELLECTUAL ASSET / 03R
Production Recipes
A reusable path from brief to delivery: planning, asset slots, tool and model routing, constraints, checks and revisions. Recipes make creative quality repeatable without flattening it.
INTELLECTUAL ASSET / 04L
Learning signals
Trajectories, critiques, corrections, preferences, interventions and accepted outputs—captured during project-scoped expert production, ready to improve models and agents.
DATA PRODUCTION ENVIRONMENT / EXPERT-OPERATED
The environment behind the data.
One2X produces datasets inside a dedicated expert content environment—not through detached labeling. Briefs, approved inputs, models, tools, edits, reviews and accepted outputs remain linked, so every production run can become traceable media, lineage, judgement and training signals.
01 / PRODUCTION SURFACESEIGHT MODES · ONE DATA ENVIRONMENT
01
Prompt to VideoGENERATE · DIRECT
02
Infinite CanvasMAP · COMPOSE
03
Video EditorCUT · TIME · REVISE
04
DocumentsWRITE · STRUCTURE
05
TablesORGANIZE · SCORE
06
WhiteboardPLAN · REVIEW
07
WebRESEARCH · PUBLISH
08
ScriptsAUTOMATE · CONNECT
02 / HOW THE DATA IS MADETASK → MEDIA → JUDGEMENT → SIGNALS
01 / DEFINEScoped taskSpecify the content capability, approved inputs, constraints and acceptance criteria.
02 / PRODUCEContent variantsExperts operate models, tools, assets and Recipes to create controlled outputs.
03 / JUDGEExpert signalsCritique, revision, preference and acceptance stay attached to each run.
04 / PACKAGETraining-ready outputMedia, lineage, labels and learning signals are quality-checked and delivered.
ONE2X / DATA PRODUCTION ENVIRONMENTEXPERT-OPERATED · LIVE SYSTEM
INPUTS06
V
Source mediaVIDEO · IMAGE · AUDIO
B
Creative briefINTENT · CONSTRAINTS
R
RecipeMODELS · TOOLS · STEPS
C
ReferencesSTYLE · BRAND · TASTE
SCENE 03 / CAMPAIGN CUTMake the idea move.00:08.12 / 00:18.00
EXPERT JUDGEMENTLIVE
Composition PASS
Message hierarchy holds across the vertical frame.
Pacing REVISE
Hold the product beat +08 frames before CTA.
Continuity PASS
Identity, palette and spatial logic remain stable.
Preference A / B
Variant B accepted. Critique stays attached to the run.
OUTPUTEDITAUDIOFEEDBACK
04 / START WITH THE MEDIUM
Start with the content capability your model needs.
We define what good looks like in that medium, choose the right production environment, and deliver the data in a form your training stack can consume.
01
Video
What we can produceMulti-scene stories, explainers, ads, character-led content and longer-form video.
Data we can deliverFinished clips, shot and scene labels, prompt–video pairs, edit traces, critiques, revisions, preferences and evaluation sets.
02
3D / Geometry
What we can produceScene, action and camera-directed assets; white-model to finished visual; controlled style transfer.
Data we can deliverGeometry–target pairs, action and camera metadata, control passes where available, retention labels and expert QA.
03
Agentic editing
What we can produceAI video edits that plan, select, call tools, compose and revise across longer tasks.
Data we can deliverPlanning trajectories, tool calls, intermediate edits, human corrections, preferences, cost and intervention data.
04
Editorial
What we can produceArticles, images and layouts designed as one coherent piece—not three separate generations.
Data we can deliverOutline, research, image-selection and layout trajectories; critique, revision, preference and final acceptance.
From task definition to a production-ready data service.
Each engagement is shaped by a model capability and a content medium. The output can be a dataset, an evaluation service, training signals or a continuously refreshed API.
Specify the task and good
Capability gap, medium, references, constraints, rubric and acceptance criteria.
OUTPUT / TASK SPEC + EVALUATION RUBRICRun the content environment
Creators, models, tools and Recipes generate content, traces and expert judgement.
OUTPUT / CONTENT + TRACE + JUDGEMENTDeliver and verify
Structure, label, quality-check and package the signals for training or evaluation.
OUTPUT / DATASET · EVAL · SIGNALS · APISERVICE EXTENSION / ENTERPRISE PLATFORM
Build your own content flywheel.
Data delivery is the default: One2X operates the environment and delivers the resulting datasets, evaluations and training signals. For ToB customers who want to internalize the capability, we can also license and deploy the platform around your content medium, models, tools and quality system.
YOUR TASKS → CREATE → JUDGE → CAPTURE → IMPROVE → REPEATYOUR CONTENT · YOUR RECIPES · YOUR PROPRIETARY SIGNALS
01 / PRIVATE DEPLOYMENTCUSTOMER-CONTROLLED
Run inside your boundary.
Deploy an isolated environment around private assets, models, toolchains, access controls and security requirements.
FIT / SENSITIVE ASSETS · PRIVATE MODELS · COMPLIANCE02 / CUSTOM WORKBENCHTOB CO-BUILD
Encode your production grammar.
We co-design complex, medium-specific creation consoles that connect your models, tools, review gates, Recipes and expert judgement.
FIT / MULTI-TOOL · LONG-HORIZON · HUMAN-IN-THE-LOOP03 / HOSTED SAASON REQUEST
Start with hosted access.
Teams that need a ready creation workspace can contact us to discuss enterprise access, configuration and workflow fit.
DISCUSS PLATFORM ACCESS↗We know content.We manufacture its data.
One2X Data is a data service built by an AI Native content company. Separate expert production environments create content-native datasets, evaluations and training signals at scale.
CONTACT US↗