Claru vs Human Archive: Agree on the Signals Before the Volume
Last updated: November 11, 2026. If anything here is inaccurate, email [email protected].
The key difference to validate
- Evaluate Claru for household/workplace activity coverage with explicit task quotas and acceptance gates.
- Evaluate Human Archive when HA-Multi's stereo depth, tactile gloves, body IMUs and wrist cameras are required recorded channels.
- Human Archive also describes HA-Ego RGB/wrist capture. It is not exclusively a sensor-heavy offer.
- Inspect recorded versus derived channels, temporal alignment and complete task execution before commissioning volume.
Two related offers, not identical sample contracts
Human Archive describes custom wearable rigs, collection operations and QA/alignment from raw capture to training data. Its HA-Multi includes egocentric vision, stereo depth, tactile gloves, body IMUs and wrist cameras; HA-Ego is an RGB/wrist-camera offering. Its website also invites catalog and deployment enquiries. Capture and QA description · Current offer
Claru's household case documents chest/waist video, task/country quotas and acceptance gates; its workplace case covers separate commercial environments. Its caption-verification case describes human review of temporal action captions on human egocentric video for a robotics research program. That is annotation QA, not robot-trajectory capture or native sensor measurement. Household · Workplace · Caption verification
The buyer's task is to distinguish the measurement contract from the activity contract. Which channels must be native recordings? Which actions and settings must be covered? Neither broad program scale nor a list of sensors proves the proposed batch meets both requirements.
When Human Archive merits evaluation
Sensor-native human capture
HA-Multi explicitly lists depth, tactile, body and wrist channels alongside egocentric vision. Ask for a complete, aligned sample of the streams your model needs. Source
A distinct RGB/wrist option
HA-Ego is described separately from HA-Multi. Do not assume every catalog subset has all the same sensors or that sensor-heavy capture is the only option. Source
Capture-to-data operations
Human Archive says it covers custom rigs, collection operations and QA/alignment. Confirm what is included in the proposed dataset or deployment scope. Source
When Claru's evidence is relevant
Household activity quotas
The household case records a versioned task taxonomy, quota tracking and acceptance checks. Use it to discuss how your required activity distribution will be controlled. Source
Commercial workplace coverage
The workplace program supplies separate evidence across real commercial activity. Match the specific setting and task, not a pooled hours total. Source
Human-video action-caption verification
The case documents human verification of temporal action captions on human egocentric video for a robotics research program: does the described action occur in the marked interval, performed by the person in frame, with the objects named? This is annotation QA, not evidence of robot-trajectory capture. Source
Claru vs Human Archive: define the sample contract
| Dimension | Human Archive | Claru |
|---|---|---|
| Public capture reference | HA-Multi and HA-Ego human datasets with different channel sets. Source | Chest/waist household footage and separate commercial workplace collection. Household · Workplace |
| Recorded-channel reference | HA-Multi: egocentric vision, stereo depth, tactile gloves, IMUs and wrist cameras. Source | Household case documents video capture; scope any additional required sensor channel explicitly. Source |
| Quality evidence | Company-authored description includes QA/alignment. Inspect the proposed aligned streams. Source | Household acceptance gates; separate human-video action-caption verification case. Capture QA · Caption QA |
| Buyer acceptance question | Do required streams, units, timestamps and alignment survive your loader? | Do task coverage, camera framing and agreed annotations meet the brief? |
Share your tasks, environments, required channels and acceptance requirements.
Request a DatasetEvaluate Human Archive when…
- You require its documented sensor combination and want to inspect native streams together.
- Its HA-Ego or HA-Multi catalog distribution matches your tasks, or you want to discuss deployment scope.
Evaluate Claru when…
- You need a quota-driven household/workplace video program with accepted-output rules.
- You need human verification of action captions on human egocentric video as a separately scoped service.
Accept the channels and the activity separately
A two-provider pilot should expose mismatches before volume is commissioned.
Specify native measurements
List mandatory recorded channels, units, time bases and calibration. Keep derived estimates separate; inferred depth is not automatically a substitute for recorded stereo depth or tactile readings.
Review a full task
Inspect start-to-finish activity, occlusion, temporal alignment and annotation boundaries in your own loader. Apply the same task completion and channel-completeness rules to both proposals.
Agree the scope to scale
Define task/environment quotas, accepted units, rejections, rights and delivery format. Keep optional caption verification or deployment work visible in the quote instead of hiding it in a footage rate.
Claru program evidence for the activity brief
Inspect the underlying scope
Expand the shortlist or examine capture and verification separately.
Claru vs Human Archive: buying questions
Is Claru video equivalent to HA-Multi?
No equivalence is established here. If your model needs recorded depth, tactile, body-motion or wrist-camera channels, specify them and inspect an actual sample. Ordinary video or derived estimates cannot be assumed to preserve those native measurements.
Does Human Archive only offer multimodal sensor data?
No. Its company-authored description includes HA-Ego with egocentric RGB and wrist cameras as well as HA-Multi. Confirm which subset and channels the proposed purchase includes. Source
Does Claru's caption-verification case prove robot-data collection?
No. The case study verifies temporal action captions on human egocentric video for a robotics research program. Reviewers check whether the action occurs in the marked interval, performed by the person in frame, with the objects named. It documents annotation QA, not robot-trajectory capture or native robot sensor measurements.
How should I compare a quote?
Match required channels, task distribution, accepted-output units, annotation, rights and delivery format. Separate any capture, verification and deployment work. No comparable public numeric buyer rate was established from the offering pages reviewed November 11, 2026.
Scope the activity and verification work
Share your tasks, environments, required channels and acceptance requirements.