FULL-TIME TEAMS FOR COMPLEX PHYSICAL AI DATA

Training-ready data pipelines for physical AI.

HumanRelay turns real-world activity into training-ready data for physical AI models.

Physical AI data pipelines depend on full-time teams who understand what is required and can deliver it consistently at the accuracy and quality levels required to produce high-performing models. To ensure this, one-third of our employees hold technical degrees, like engineering. This becomes increasingly important as capture devices and enterprise datasets become more complex.

Start a Pilot Request Sample Data
7,000+ hrs / day
capture trajectory by the end of September
500+ million
annotations delivered
1,500+ FTE
full-time, managed professionals
15 offices
across seven states in India
HumanRelay physical AI data pipeline connecting managed human teams with robot and AI systems
Fig. 1 — Human data infrastructurePhysical AI
WHY MANAGED TEAMS MATTER

High-performing models depend on data that is captured, reviewed, and delivered to the same quality bar every time — not intermittent throughput from an anonymous crowd.

Technical fluencyTeams learn the capture devices, task logic, data requirements, and enterprise environments they work in.
Calibrated qualityFull-time teams retain their training and calibration, so the quality bar improves instead of resetting with each task.
End-to-end accountabilityNamed employees own the work from collection through delivery, making quality consistent and traceable.
THE STACK

One partner. The whole human stack.

Everything your agents and robots need humans for — from pretraining data to live judgment to deployment fallback. Four products, one workforce, one API.

01 /

Capture

Egocentric data collection

First-person 4K video of real people doing real work — hands, tools, homes, kitchens, warehouses.

  • • +3,000 hours captured daily, growing fast
  • • Head-mounted 4K rigs with synced audio + IMU
  • • Scripted task coverage or naturalistic capture
  • • Consent-first, full provenance metadata
Explore Capture →
02 /

Annotate

Labeling & evaluation

Delivered by IndiVillage, our sister company — 500M+ annotations for global AI teams.

  • • Video, sensor, and 3D/pose labeling
  • • Taxonomy & ontology design
  • • RLHF preference ranking & rubric evals
  • • Dual-pass QA at 99%+ accuracy
Explore Annotate →
03 /

Judge

Human-in-the-loop API

Five primitives your agents call when they need a human: Classify, Judge, Extract, Escalate, Resolve.

  • • One REST API or MCP server
  • • Relay decomposes complex questions into priced binaries
  • • Sub-minute responses, full audit trail
  • • From $0.50 per call
Explore Judge →
04 /

Operate

Teleoperation

When a deployed robot hesitates, a trained operator takes over in seconds.

  • • 24/7 operator benches, latency-tiered
  • • Safety confirmation before risky actions
  • • Full intervention trace returned via API
  • • Every intervention becomes training data
Explore Operate →
THE FLYWHEEL

Every intervention is a labeled demonstration.

Your hardest moments become your model's best data. A robot stalls — or an agent hits a question it shouldn't guess at. Our people resolve it live. The resolution comes back annotated and training-ready. Next quarter's model escalates less. That loop is the product.

"Every human knows not to drive into a flooded road. No dataset does."

We call it the common-sense gap. Models learn from what got recorded. The real world keeps producing what didn't: the flooded road, the cones that contradict the lane paint, the customer photo that's 60% wasp. Too rare, too new, too ambiguous for any training set — the long tail of the real world, and deployment finds every inch of it.

The flywheel closes that gap. When your robot or agent meets a situation no dataset covered, a person who has lived it resolves it in seconds — and the resolution returns as a labeled demonstration. The edge case your fleet hits this morning is training data by tonight.

1

Capture

Egocentric human video at scale seeds the model.

2

Train

Annotated, taxonomy-aligned data goes into pretraining.

3

Deploy

Your robots and agents ship before the model is perfect — safely.

4

Escalate

Edge case hit: the robot or agent requests a human via API.

5

Intervene

Judgment call or full teleop — resolved in seconds.

6

Annotate

The intervention returns as a labeled demonstration.

↺ Back into training — fewer escalations every cycle
Seconds
from escalation to a trained human on task
100%
of interventions returned as training-ready data
One vendor
from pretraining data to deployment fallback
01 — CAPTURE

3,000+ hours of physical AI data. Every day.

Egocentric data collection. First-person 4K video of real people doing real work — hands, tools, homes, kitchens, warehouses.

Robots learn manipulation from human hands. Across 15 offices and seven states, we run managed capture centers where trained collectors — on staff, not crowdsourced — record first-person video of real tasks: cooking, cleaning, folding, assembly, picking, repair.

You direct the taxonomy. We deliver the hours — raw, curated, or fully annotation-ready through our own labeling teams.

  • →Already running. Over 3,000 hours of 4K capture per day, with a path to 7,000+ hours per day by the end of September.
  • →Directable. Scripted task programs against your taxonomy, or naturalistic full-day capture.
  • →Clean provenance. Consent-first collection, full metadata, PII scrubbing.
  • →One pipeline. Capture and annotation under one roof — no vendor hand-off.
Request a Sample DatasetEgocentric data collection

Capture Spec

RigHead-mounted 4K @ 30/60 fps
StreamsVideo · audio · IMU (gaze optional)
CoverageScripted tasks or naturalistic days
Throughput100s of hours / day, growing
ProvenanceConsent-first · metadata · PII-scrubbed
DeliveryRaw · curated · annotation-ready
Custom programs: need a specific environment, tool set, or demographic mix? We stand up dedicated capture programs against your spec.
02 — ANNOTATE

Annotation that's already proven at scale.

We don't claim a labeling capability — we point at one. Our annotation operation is run with our sister company, IndiVillage.

Powered by IndiVillage

Sister company

IndiVillage has delivered 500M+ annotations for global AI teams from managed delivery centers — salaried professionals working in offices, trained per project, retained for years. Its impact-sourcing model builds tech careers in communities that rarely get them: quality your ML team can verify, and a workforce story your stakeholders can be proud of.

Capabilities

09
Egocentric video labeling
3D pose & hand tracking
Object & affordance tagging
RLHF preference ranking
Agent trace evaluation
Multi-turn conversation rating
Content moderation
Document extraction
Custom taxonomy & ontology design
QA built in: dual-pass review, gold-set sampling, and calibration sessions hold delivered accuracy above 99%. Pilots start from as few as 10 hours.
03 — JUDGE

An API for human judgment.

Five primitives your agents and robots call when they need a human. One REST API or MCP server. Sub-minute responses with full audit trails.

The Primitives

Classify
Binary or multi-class categorization by trained humans.
Judge
Pairwise comparison, RLHF ranking, subjective evaluation.
Extract
Structured data from documents, images, or video.
Escalate
Route high-risk or low-confidence cases to specialists.
Resolve
End-to-end human resolution with rationale and audit trail.
Basic
$0.50 / call
Complex
$1.00 / call
Expert
$2.50 / call

Relay: complex questions, priced binaries.

Relay is the intelligence layer between your agent and our humans. It decomposes any question into binary decisions, routes each to the right tier, runs them in parallel, and reassembles the answer — with the full trace returned for your learning loop.

  • →Direct first. An image and a question — "is this safe to drive through?" — go straight to one human, at one price.
  • →Decompose when it earns it. Genuinely multi-part questions split into binaries that run in parallel across workers.
  • →Reassemble. Answers compose into a final response with reasoning.
"Your delivery robot sends a camera frame" → "Is this safe to drive through?"
Relay routes
One trained human, one look at the frame — Complex $1.00
"No. Water spans the road and cars are turning back — reroute."
direct · 1 human · 19 seconds$1.00 total
Call the API

One REST API or MCP server.

Judgment on tap for production agents. Five HITL primitives behind one API or MCP server — moderation calls, RLHF, evals, escalation.

Sub-minute human answers for the moments your agent shouldn't guess. From $0.50 a call.

Python
from humanrelay import HumanRelay

hr = HumanRelay(api_key="hr_live_...")

result = hr.judge(
    content={"a": response_a, "b": response_b},
    rubric="Which response is more helpful?",
    tier="expert",
)

print(result.verdict)    # "Response A"
print(result.rationale)  # "Response A provides..."
04 — OPERATE

A human hand on the wheel, in seconds.

Deployed fleets meet the long tail: a jammed gripper, an ambiguous object, a customer at the door. Operate puts trained teleoperators behind your robots around the clock.

The robot requests help. We confirm safety. An operator takes over. Control hands back. You get the resolution — and the recording, annotated, as a demonstration for your next training run.

  • →Staffed benches, 24/7. Operators on shift in our centers — not on-call gig workers.
  • →Latency-tiered routing. Standby SLAs matched to task risk.
  • →Safety gate. A $0.50 confirmation binary before any consequential physical action.
  • →Data exhaust included. Full intervention trace plus annotated demonstration, returned via API.
Discuss a Teleop Pilot
No fleet yet? We also run scripted teleoperation programs purely for data collection — on your rigs or ours.

Intervention Lifecycle

1
Escalation
Robot calls the API with context
2
Safety check
Confirmation binary, sub-minute
3
Session
Operator takes control (low-latency)
4
Handback
Robot resumes autonomous operation
5
Return
Trace + annotated demonstration
THE WORKFORCE

A managed, skilled workforce makes the difference.

HumanRelay employs over 1,500 full-time professionals, approximately one-third with engineering or other technical degrees. Gig platforms churn anonymous workers through your data; we put named, trained, salaried professionals on your project — in offices we run.

Quality

Same people on your project every day. Project-specific training, calibration sessions, dual-pass QA. That's how you hold 99%+ accuracy and 98% client retention — numbers a revolving crowd can't reach.

Security

Access-controlled floors, NDAs, managed devices, SOC 2 compliance. Your pretraining data and customer footage never touch an anonymous crowd — every person who sees it is accountable by name.

Impact

Through IndiVillage's impact-sourcing model, this work builds careers in communities that rarely get them. Procurement will like the security posture. Your board will like the story.

GLOBAL FOOTPRINT

Global reach, rooted in India.

Headquartered in London, HumanRelay operates 15 offices across seven Indian states. Together, those states are home to nearly 700 million people and around 7 million registered businesses — a breadth of ground-level coverage few data providers can replicate. Combined with our in-house annotation capabilities, this footprint enables HumanRelay to deliver training-ready data pipelines for physical AI models at enterprise scale. Beyond India, a global field-collection partner network extends that reach worldwide. Capture and labeling stay under one accountable system, run by salaried teams rather than a gig crowd.

OPERATING STATES
7
  • Andhra Pradesh
  • Jharkhand
  • Karnataka
  • Madhya Pradesh
  • Maharashtra
  • Rajasthan
  • Uttar Pradesh
Reach
~700M
People
~7M
Registered businesses

Combined state totals — a depth of presence few data providers can replicate.

London
Global headquarters

Strategy, partnerships, and client teams based in the UK.

15
Offices across 7 states

Salaried, in-office delivery teams across seven Indian states — capture and annotation under one roof.

Global
Capture partner network

Beyond India, a worldwide network of partners for on-location data collection.

WHO IT'S FOR

Built for the people building agents and robots.

For agent builders

Judgment on tap for production agents. Five HITL primitives behind one API or MCP server — moderation calls, RLHF, evals, escalation. Sub-minute human answers for the moments your agent shouldn't guess. From $0.50 a call.

Start a pilot →

For humanoid companies

Deploy before the model is perfect. Pretraining ego-video at scale, evaluation pipelines, and a 24/7 teleop fallback behind every unit in the field — so the fleet ships now and improves monthly.

Start a pilot →

For frontier labs

Data you can put in the model card. Diverse, consented, provenance-clean human data for VLA and world models. RLHF, rubric evals, and red-teaming by trained specialists — not a marketplace.

Request sample data →

For investors

The bottleneck is the business. Embodied AI's constraint isn't compute — it's human data. Ask for the memo: traction, the IndiVillage moat, and the intervention flywheel.

Request the memo →
PRICING

Four products. Simple meters.

Pay for delivered hours, labeled assets, resolved calls, or completed interventions. No platform fees, no seat licenses.

Capture

per data-hour

Program-based egocentric capture, priced per delivered hour by spec.

  • • Pilot batches to standing daily capture
  • • Raw, curated, or annotation-ready
  • • Custom environments & taxonomies

Annotate

per asset / per hour

Labeling and evaluation with QA included, via IndiVillage delivery centers.

  • • Pilots from as few as 10 hours
  • • Volume pricing at scale
  • • Dedicated teams for standing work

Judge

$0.50 – $2.50 per call

Human judgment by API: Basic $0.50 · Complex $1.00 · Expert $2.50.

  • • Sub-minute response targets
  • • Relay decomposition for complex asks
  • • Full audit trail on every call

Operate

standby + per intervention

Teleoperation SLAs tiered by latency and risk, plus per-intervention pricing.

  • • 24/7 staffed coverage
  • • Safety-gated control sessions
  • • Annotated demonstrations included
Volume discounts at scaleSLA guarantees availablehello@humanrelay.com for a quote
FAQ

Questions, answered straight.

One partner for the human side of AI — software agents and embodied AI alike: egocentric data capture, annotation and evaluation (run with our sister company IndiVillage), a real-time human-judgment API your agents call when they shouldn't guess, and teleoperation for deployed robot fleets. One workforce of 1,500+ managed professionals powers all four.

Put 1,500 managed professionals behind your agents and robots.

Capture, annotation, judgment, teleoperation — scoped as a pilot this week.

Start a Pilot
No commitment requiredPilot results in 48 hoursSOC 2 compliant
Or email us directly: hello@humanrelay.com