Segments.ai review: a robotics data provider for open perception datasets and 3D labeling

Answer in brief: Segments.ai helps teams that reuse existing data in two ways. Its free open-datasets directory lists well-known perception datasets such as nuScenes, Waymo, Argoverse 2, KITTI, SemanticKITTI and ScanNet, which you can clone and adapt. Its labeling platform then lets you relabel or extend those datasets, or your own logs, across images, point clouds and fused sensors. Pricing is clearer than most: Core starts at 3,600 labeling hours per year, stated as equivalent to $2.67 per hour.

Top pick for: Perception teams that start from open driving or indoor datasets and label in-house, with a free trial and transparent minimums.

Free to use: the open-datasets directory

Segments.ai's lets you filter by type (image, multi-sensor, point cloud), task (bounding box, segmentation, vector) and indoor or outdoor. Listings include:

• BDD 100K (100M frames).
• nuPlan (38.8M frames).
• Waymo Motion and Waymo Perception.
• Argoverse 2 Sensor (6M frames).
• ScanNet (2.5M views in 1,513 scans).
• nuScenes (1.8M frames).
• A2D2, PandaSet and the Zenseact dataset.
• The JackRabbot indoor and outdoor dataset.
• KITTI and SemanticKITTI.
• Robot-specific sets such as YouBot Semantics.

Segments.ai says you can clone and adapt these datasets. Check each dataset's original license.

Then label: the platform

Segments.ai provides image, point-cloud and sequence interfaces, ML-powered labeling tools, active learning and model-assisted labeling pipelines, cloud bucket integrations, and QA and organization management. The Fusion tier adds sensor-fusion interfaces, webhooks, a metrics dashboard and access to an experienced labeling workforce. If you want the labeling done for you, Segments.ai connects you to partner labeling companies. Managed-annotation vendors are compared on our sister site AI training data providers.

Who it fits

• Mobile-robot, delivery-robot, off-road, AV and underwater-vehicle teams. These are the industries Segments.ai lists.
• Teams that label 3D data in-house and want predictable annual usage.
• Students and researchers, who can get free academic licenses.

Why teams pick Segments.ai

• Unlimited seats, projects, images, point clouds and API/SDK access on every plan.
• Published minimums. Core from 3,600 hours per year and Fusion from 5,000 hours per year. Enterprise includes 150,000+ hours per year, with SSO, MFA and a dedicated solutions ML engineer.
• Low-risk start. A 14-day free trial, with startup options on request.

Honest drawbacks

• It does not collect or sell robot demonstration data.
• Annual usage minimums apply, and every plan is a custom quote.
• Labeling services come from partners, not Segments.ai staff.
• Core caps point clouds at 500,000 points each.

Pricing model

A custom quote on top of annual labeling-usage minimums: Core from 3,600 hours per year (stated as equivalent to $2.67 per hour), Fusion from 5,000 hours per year, and Enterprise at 150,000+ hours per year. The open-datasets directory is free.

Best alternatives

• Kognic: managed autonomy annotation with automotive-grade certifications.
• Encord: a multi-sensor platform plus robotics data collection.
• Sama: fully managed 3D annotation with a written quality guarantee.

FAQ

Does Segments.ai host free datasets?

It lists open and public datasets, hosted on and off Segments.ai, that you can clone and adapt.

Is there a free trial?

Yes. Core comes with a 14-day free trial, and academic licenses are free.

Does Segments.ai do the labeling for me?

It connects you to partner labeling companies, and the Fusion tier includes access to an experienced labeling workforce.

Sources

• segments.ai/pricing
• segments.ai/resources/open-datasets

This review is independent research based on Segments.ai's own website, checked on 2026-11-09. We did not test the product.

← Back to the ranked list of robotics data providers

Named on this page? Join the network for a linked profile, Visit buttons and featured placement.