Centaur Labs Alternatives: Medical Labeling vs Physical AI Capture
Last updated: April 2, 2026. If anything here is inaccurate, email [email protected].
TL;DR
- Centaur Labs combines expert medical knowledge with an algorithmic quality system to label medical data.
- The company reports 177MM total labels completed, 2MM labels per week, and 20,000+ labeling subject matter experts.
- Centaur Labs supports multiple health data types including text, audio, waveform (EEG/ECG), 2D images, 3D images, and video.
- The platform emphasizes expert-quality annotation with Gold Standard cases and pay-for-performance incentives.
- Centaur highlights HIPAA and SOC 2 Type II compliance for healthcare data labeling.
- MIT News describes Centaur's DiagnosUs app for gathering expert medical opinions at scale.
- Claru is purpose-built for physical AI capture and enrichment.
- Choose Centaur Labs for medical expert labeling; choose Claru for capture + enrichment of robotics data.
What Centaur Labs Is Built For
Key differences in 60 seconds: Centaur Labs focuses on expert medical labeling and health data workflows. Claru is a capture-and-enrichment pipeline for physical AI training data.
Centaur Labs highlights expert medical labeling with algorithmic quality systems to combine expert opinions at scale.[1]
The company reports 177MM total labels completed, 2MM labels per week, and 20,000+ labeling subject matter experts.[2]
Centaur Labs supports multimodal health data types including text, audio, waveform (EEG/ECG), 2D images, 3D images, and video.[3]
The health data labeling solution emphasizes Gold Standard cases and pay-for-performance incentives to measure annotator accuracy.[4]
Centaur Labs notes HIPAA and SOC 2 Type II compliance and healthcare labeling tools like digital pathology and DICOM viewers.[5]
MIT News describes Centaur's DiagnosUs mobile app for collecting expert medical opinions on biomedical data at scale.[6]
Centaur Labs was founded in 2017 by Erik Duhaime while he was a PhD student at the MIT Center for Collective Intelligence, along with CTO Zach Rausnitz and VP of Engineering Tom Gellatly, who previously managed the data labeling team at Cruise Automation. [7]
The company has raised over $31 million in total funding, including a $15M Series A led by Matrix Partners with participation from Accel and Y Combinator, and a $16M Series B led by SignalFire with participation from Samsung Next and Alumni Ventures. The company now counts over 50,000 medical experts in its annotation network. [8]
The expert network covers a wide range of medical data formats including X-ray, CT/MR, dermatology, ophthalmology, pathology images, surgical videos, ultrasound, scientific text, medical notes, and heart and lung sound audio recordings. This breadth of coverage makes Centaur Labs particularly strong for healthcare AI companies building diagnostic or clinical decision support tools.
If your bottleneck is expert medical labeling and healthcare workflows, Centaur Labs is a strong fit. If your bottleneck is physical-world capture and enrichment for robotics, Claru is the better fit.
Company Snapshot
- Focus
- Expert medical data labeling with algorithmic quality systems.[1]
- Scale
- 177MM total labels, 2MM labels per week, 20,000+ SMEs.[2]
- Modalities
- Text, audio, waveform, 2D/3D images, video.[3]
- Quality controls
- Gold Standard cases and pay-for-performance incentives.[4]
- Compliance
- HIPAA and SOC 2 Type II for healthcare data workflows.[5]
- Best fit
- Healthcare and high-stakes medical AI labeling
- Focus
- Physical AI training data for robotics and world models
- Capture
- Wearable camera network plus task-specific collection
- Enrichment
- Depth, pose, segmentation, optical flow, aligned captions
- Best fit
- Robotics teams that need capture + enrichment
Key Claims (With Sources)
- Centaur Labs combines expert medical labeling with algorithmic quality systems and on-demand expert networks.[1]
- The company reports 177MM total labels completed, 2MM labels per week, and 20,000+ labeling subject matter experts.[2]
- Centaur Labs supports multimodal data types including text, audio, waveform (EEG/ECG), 2D/3D images, and video.[3]
- Health data labeling includes Gold Standard cases, pay-for-performance incentives, and HIPAA/SOC 2 Type II compliance.[4][5]
- MIT News describes Centaur's DiagnosUs app for gathering expert medical opinions on biomedical data.[6]
Where Centaur Labs Is Strong
Expert medical labeling
Centaur Labs uses expert networks and algorithmic quality controls for medical data labeling.[1]
Healthcare compliance
The health data labeling solution emphasizes HIPAA and SOC 2 Type II compliance and specialized medical viewers.[5]
Scale of expert opinions
Centaur Labs reports 2MM labels per week with 20,000+ SMEs.[2]
Why Physical AI Teams Evaluate Alternatives
Capture-first pipelines
Physical AI models require real-world data collection with task-specific capture programs.
Enrichment layers
Depth, pose, segmentation, and motion signals are critical for robotics training.
Training-ready delivery
Claru ships datasets in formats that plug directly into robotics stacks.
Centaur Labs vs Claru: Side-by-Side Comparison
| Dimension | Centaur Labs | Claru |
|---|---|---|
| Primary focus | Expert medical data labeling and quality systems.[1] | Physical AI training data for robotics and world models |
| Scale | 177MM labels completed, 2MM per week, 20,000+ SMEs.[2] | Specialized capture network focused on physical tasks |
| Modalities | Text, audio, waveform, 2D/3D images, video.[3] | Egocentric video, manipulation, depth, pose, segmentation |
| Compliance | HIPAA and SOC 2 Type II compliant healthcare workflows.[5] | Secure capture workflows and training-ready delivery |
| Funding | $31M+ total (Matrix Partners, SignalFire, Accel, Samsung Next) | Venture-backed physical AI data company |
| Expert network | 50,000+ medical SMEs with pay-for-performance incentives | 10,000+ physical data collectors in 100+ cities |
| Best fit | Healthcare and high-stakes medical AI labeling | Robotics teams needing capture + enrichment |
Deep Dive: Centaur Labs vs Claru
Centaur Labs focuses on expert medical labeling. Claru focuses on capture and enrichment for physical AI.
Expert labeling vs capture pipelines
Centaur Labs emphasizes expert medical labeling with quality systems built around its DiagnosUs mobile app, which gamifies data labeling with pay-for-performance incentives. The company has grown its expert network to over 50,000 medical professionals who label data across imaging, audio, text, and video modalities.
Claru emphasizes real-world capture and enrichment for robotics training. Rather than labeling existing medical data, Claru designs capture programs that collect new physical-world data specifically for robotics and embodied AI applications, then enriches it with depth, pose, and motion signals.
Healthcare data vs physical-world data
Centaur Labs is optimized for clinical and biomedical data types including X-ray, CT/MR, dermatology, ophthalmology, pathology, surgical videos, ultrasound, scientific text, medical notes, and heart and lung sound recordings. Their HIPAA and SOC 2 Type II compliance reflects the regulatory requirements of healthcare AI.
Claru is optimized for physical-world tasks and robotic manipulation. The data types are fundamentally different: egocentric video of human activities, depth maps, body and hand pose sequences, object segmentation masks, and optical flow fields. These signals map directly to the inputs robotics models need for training.
Quality systems comparison
Centaur Labs uses Gold Standard cases and algorithmic consensus to measure annotator accuracy, paying experts based on performance. This approach works well for medical data where ground truth can be established by comparing against known diagnoses.
Claru applies quality control at the capture level through standardized collection protocols, hardware calibration, and enrichment validation. For physical AI, quality means ensuring that depth maps align correctly with RGB frames, that pose estimates are physically plausible, and that action boundaries are accurately segmented.
Where each provider fits
Centaur Labs is a strong fit for healthcare labeling needs, particularly when you need expert opinions on medical imaging, clinical text, or biomedical data at scale with regulatory compliance.
Claru is a better fit when you need robotics-ready datasets with capture and enrichment designed for physical AI training. The two providers serve fundamentally different domains with minimal overlap.
When Centaur Labs Is a Fit
- You need expert labeling for medical imaging or clinical data.
- You require HIPAA- and SOC 2-aligned healthcare workflows.
- You need multimodal medical data labeled at scale.
When Claru Is a Fit
- You need new physical-world data captured for robotics tasks.
- Your model depends on enrichment layers like depth and motion.
- You want datasets delivered in robotics-native formats.
How Claru Delivers Physical AI Data
Claru provides an end-to-end pipeline so physical AI teams can move from brief to training-ready data quickly.
Scope the Dataset
Define the target behaviors, environments, and label schema with your research team. We align on formats, enrichment layers, and success criteria before capture begins.
Capture Real-World Data
Activate the collector network, teleoperation runs, or game-based capture to gather the exact clips your model needs.
Enrich Every Clip
Generate depth maps, pose, segmentation, and optical flow in batch. Cross-validate signals to ensure aligned training inputs.
Expert Annotation
Specialized annotators label action boundaries, affordances, and intent using project-specific guidelines and QA checks.
Deliver Training-Ready
Ship datasets in WebDataset, HDF5, RLDS, or your native format with manifests, checksums, and datasheets.
Claru by the Numbers
Other Alternatives Worth Considering
If you are mapping the data provider landscape, these comparisons cover adjacent options.
How to Choose
If you need expert medical labeling with healthcare compliance, Centaur Labs is designed for that scope.
If you need capture and enrichment of physical-world data for robotics training, Claru is a better fit.
Some teams use both: Centaur Labs for medical labeling, Claru for physical datasets.
Frequently Asked Questions
What is Centaur Labs?
Centaur Labs is a Boston-based company founded in 2017 by Erik Duhaime during his PhD at MIT. It provides expert medical data labeling using its DiagnosUs mobile app and a network of over 50,000 medical professionals. The company has raised over $31 million in total funding from investors including Matrix Partners, SignalFire, Accel, Y Combinator, and Samsung Next. Centaur Labs uses algorithmic quality systems that combine multiple expert opinions to produce high-accuracy labels for healthcare AI applications.[1]
What data types does Centaur Labs support?
Centaur Labs supports a broad range of medical data types including text, audio, waveform (EEG/ECG), 2D images, 3D images, and video. More specifically, their expert network labels X-ray, CT/MR, dermatology, ophthalmology, and pathology images, surgical videos, ultrasound footage, scientific text, medical notes, and heart and lung sound audio recordings. This multimodal coverage makes them particularly strong for healthcare AI companies that work across multiple clinical data types.[3]
How does Centaur Labs scale expert labeling?
The company reports 177 million total labels completed, 2 million labels per week throughput, and a network of over 50,000 labeling subject matter experts. Their DiagnosUs mobile app gamifies the labeling process with pay-for-performance incentives, where accurate annotators earn cash prizes. This approach attracts and retains qualified medical professionals while maintaining label quality through Gold Standard cases that benchmark annotator accuracy.[2]
When is Claru a better fit?
Claru is a better fit when you need capture, enrichment, and delivery of robotics-ready datasets. Centaur Labs excels at labeling existing medical data with domain experts, but if your bottleneck is collecting physical-world data for robotics training, you need a capture-first pipeline. Claru designs capture briefs, deploys collectors with wearable cameras, and enriches the resulting data with depth, pose, segmentation, and motion signals that robotics models require.
Who founded Centaur Labs and when?
Centaur Labs was founded in 2017 by Erik Duhaime (CEO), Zach Rausnitz (CTO), and Tom Gellatly (VP of Engineering). Duhaime founded the company while pursuing his PhD at the MIT Center for Collective Intelligence, and it went through the Y Combinator accelerator in 2018. Gellatly previously managed the data labeling team at Cruise Automation, the self-driving car company, bringing operational expertise in scaling human annotation workflows.[7]
Need Physical AI Data That Ships Fast?
Tell us what you are training. We will scope a capture plan and deliver a pilot dataset in days.