Skip to the content
Know your tech before you buy it. Makers: get listed from $49/mo

Segments.ai

segments.ai

  • Free plan
  • Web
  • API
7.6 Club score Visit segments.ai

What Segments.ai is

Segments.ai is a multi-sensor data labeling platform for robotics and autonomous vehicle teams. It supports labeling images and point clouds with segmentation, cuboids, bounding boxes, keypoints, polygons, and polylines. Teams can view 2D images and 3D point clouds together and project labels between sensors. For 3D sequences, Batch Mode, merged point clouds, interpolation, and automated tracking help carry object labels through sequences. Image tools include Superpixel, Autosegment, hover-and-click segmentation, and uploaded model predictions for review and correction. A Python SDK supports uploads, label downloads, dataset management, and webhooks; TensorFlow and PyTorch are named integrations. The Core plan includes cloud bucket integrations and supports point clouds up to 500,000 points each. Fusion allows point clouds and images of unlimited size and resolution. The company says it is EU based, ISO 27001 certified, and GDPR compliant. Academic researchers can request a free license for non-commercial work. Core, Fusion, and Enterprise use custom quotes; a 14-day trial is listed. Segments.ai launched in 2020 as a segmentation-focused data labeling solution.

Who it suits

It is aimed at computer vision engineers and data scientists working in autonomous robotics and driving. Teams labeling across image and point-cloud data may find its multi-sensor workflow relevant.

Good to know, and what to watch

  • Labels images and point clouds with six annotation types.
  • Projects labels between 2D images and 3D sensors.
  • Sequence tools include interpolation and automated tracking.
  • Python SDK supports dataset management and webhooks.
  • Free academic license for non-commercial research.
  • Core point clouds are limited to 500,000 points each.
  • Core, Fusion, and Enterprise pricing requires a custom quote.

Core

—

Enterprise

—

Fusion

—

Academic license

Free

Pricing on request · 14-day trial

Getting started

  1. Open the Segments.ai website
  2. Request the 14-day trial or ask for a commercial plan quote
  3. Request the academic license if doing non-commercial academic research
  4. Choose web access or connect through the API
  5. Upload images or point clouds using the web tools or Python SDK
  6. Select annotation types and export the labeled dataset

Limits to check

The free academic license is only for non-commercial academic research and requires a license request. Core supports point clouds up to 500,000 points each; Fusion includes point clouds and images of unlimited size and resolution.

The Club's review

Segments.ai is a data-labeling platform for teams working with images and 3D point clouds, especially in robotics and autonomous driving. It suits computer vision engineers and data scientists whose work spans sensors or sequences. Its strongest case is the joined 2D and 3D workflow; Core’s point-cloud size ceiling and custom-quote pricing may rule it out for some projects.

Overview

Segments.ai brings image and point-cloud labeling into one workspace, where teams can view both data types together and project labels between sensors. That makes it a more targeted choice than a general image-labeling tool for teams aligning data from multiple sensors. The platform launched in 2020 as a segmentation-focused labeling solution and is available on the web and through an API.

Teams can collaborate and export datasets in JSON, PyTorch, TensorFlow, and Hugging Face formats. The Python SDK supports uploads, label downloads, dataset management, and webhooks, giving engineering teams ways to connect annotation work to their own pipelines.

Key features

Flexible 2D and 3D annotation

Annotation options cover segmentation, bounding boxes, polygons, polylines, keypoints, and cuboids. This range suits datasets that need both object-level labels and more detailed shapes. Image support includes JPEG, PNG, and BMP.

Multi-sensor and sequence workflows

Viewing images alongside point clouds and projecting labels between sensors can help keep annotations aligned across a multi-sensor dataset. For 3D sequences, Batch Mode, merged point clouds, interpolation, and automated tracking support labeling objects over time. These tools make the platform particularly relevant to robotics and autonomous-driving data, where sequences and sensor relationships matter.

Model-assisted labeling and integrations

Image workflows include Superpixel, Autosegment, hover-and-click segmentation, and the ability to review and correct uploaded model predictions. These options can assist annotation work, though teams still need to review predictions. TensorFlow and PyTorch are named integrations, and Core includes cloud bucket integrations.

Security and enterprise access

Segments.ai states that it is ISO 27001 certified, GDPR compliant, and EU based. Enterprise adds SSO and MFA, custom security solutions, and a dedicated solutions ML engineer. Those provisions may matter to organizations with specific access and security requirements.

Pricing

Segments.ai has a free plan for non-commercial academic research and a 14-day trial. The paid plans use custom pricing, so teams should request a quote to compare cost with their expected labeling usage.

  • Academic license: free, billed 100% free for non-commercial academic research. It is for students and researchers, and requires a license request. Its non-commercial restriction makes it unsuitable for commercial projects.
  • Core: Custom quote, starting at 3,600 hours per year of labeling usage, equivalent to $2.67/hour. It includes unlimited seats, projects, images, and point clouds, but each point cloud can contain up to 500,000 points. That cap is a key constraint for teams working with larger scans.
  • Fusion: Custom quote, starting at 5,000 hours per year. It includes unlimited seats, projects, images, and point clouds, with point clouds and images of unlimited size and resolution. Priority support is available through mail, Slack, and Discord.
  • Enterprise: Custom quote, including more than 150,000 hours per year of labeling usage and unlimited seats, projects, images, and point clouds. It also adds SSO and MFA, custom security solutions, and a dedicated solutions ML engineer.

The plans all support unlimited seats and projects, while usage thresholds and data-size limits distinguish them. Core is the lower starting point for substantial labeling workloads, but teams with point clouds above its stated limit should consider Fusion or Enterprise.

Platforms

Segments.ai is available through its web platform and API. The Python SDK supports programmatic dataset and label workflows, and dataset exports include JSON, PyTorch, TensorFlow, and Hugging Face formats.

Who it's for

Computer vision engineers and data scientists in autonomous robotics and driving are the clearest fit, particularly when their projects involve both images and point clouds or labeling objects across sequences. Teams that need varied annotation types, model-assisted image tools, or programmatic access also have a strong reason to consider it. A team labeling only still images may not need its multi-sensor workflow, while projects requiring point clouds above 500,000 points per cloud exceed Core’s limit.

Pros and cons

  • Pro: Teams can view images and point clouds together and project labels between sensors, supporting coordinated multi-sensor annotation.
  • Pro: Batch Mode, interpolation, merged point clouds, and automated tracking address labeling across 3D sequences.
  • Pro: The annotation set spans segmentation, boxes, polygons, polylines, keypoints, and cuboids, with model-assisted image tools and uploaded predictions.
  • Pro: Unlimited seats and projects are included across the paid plans, and the Python SDK supports dataset and label workflows.
  • Con: Core limits each point cloud to 500,000 points, which may exclude larger scans unless a higher tier fits.
  • Con: Paid plans require a custom quote, so teams cannot compare a published plan price directly.
  • Con: The free Academic license is restricted to non-commercial academic research and requires a license request.

Alternatives

Image Annotation Software, AI Image Annotation Tools, and Data Labeling Software offer broader directories for comparing tools by category.

  • Label Studio is worth considering for teams seeking a freemium option with API, web, desktop, or self-hosted platform choices.
  • Supervisely may suit teams wanting a freemium path with a defined Community allowance of 5 GB storage, 10,000 files, and two members.
  • Roboflow is an option for teams looking for a freemium platform with a stated monthly credit allowance on its Free Tier.
  • CVAT may fit personal use and small teams seeking a free Community edition under an MIT license.
  • MakeSense.ai is a free, browser-based photo-labeling option for teams with simpler image needs.
  • Labellerr offers a free Researcher Plan with 2,500 data credits, one seat, one workspace, and up to 10 projects.
  • Draw Attention is a free option for creating one interactive image.
  • Ultralytics Platform is another freemium option for teams comparing platforms with a free plan.

Verdict

Choose Segments.ai if your team labels robotics or autonomous-driving data across images, point clouds, and sequences: its multi-sensor workflow and 3D sequence tools are the main reasons to pick it. Look elsewhere if you need a published paid price, a commercial free tier, or point clouds larger than Core allows without moving to a higher custom-quoted plan.

The verdict

Segments.ai focuses on multi-sensor annotation, with tools for both images and point-cloud sequences. Check Core’s point limit and request a quote to assess a paid plan.

Questions people ask

Is there a free option?

Yes. The Academic license is free for non-commercial academic research, for students and researchers, and requires a license request.

How is commercial pricing handled?

Core, Enterprise and Fusion are offered by custom quote. Core starts at 3,600 labeling hours per year, Enterprise includes 150,000 hours per year, and Fusion starts at 5,000 hours per year.

Which annotation types are supported?

Segments.ai supports segmentation, bounding boxes, polygons, polylines, keypoints and cuboids for images and point clouds.

Can it connect to machine-learning workflows?

Yes. The Python SDK supports uploads, label downloads, dataset management and webhooks, and TensorFlow and PyTorch are named integrations.

What platforms are available?

The product is available through web access and an API.

Is there a trial?

Yes. A 14-day trial is listed.

What the maker publishes

2D and 3D annotation
The platform supports image and point-cloud labeling with segmentation, cuboids, bounding boxes, keypoints, polygons, and polylines. source
2D annotation tools
Image labeling supports semantic, instance and panoptic segmentation plus bounding boxes, polygons, polylines, keypoints, superpixels, Autosegment and hover-and-click tools. source
3D point-cloud tools
3D workflows include cuboids, segmentation, keypoints, polygons, polylines, merged point-cloud mode, batch mode and automated tracking. source
3D sequence tools
Batch Mode, merged point clouds, interpolation, and automated tracking support labeling objects across sequences. source
API scope
The documented API covers datasets, samples, labels, labelsets and releases for programmatic annotation workflows. source
Asset upload limit
The platform's own asset-storage service limits uploaded files to 100 MB and stores them in an AWS S3 bucket in London. source
Cloud integrations
Documentation describes AWS S3 cross-account access, Azure Blob Link Service and keeping assets in customer-managed cloud buckets. source
Cloud storage
The Core plan includes cloud bucket integrations. source
Company history
Segments.ai says it launched in 2020 as a data labeling solution designed for segmentation. source
Company size and base
LinkedIn lists Segments.ai by Uber as a Leuven, Belgium software-development company with 11–50 employees and a 2020 founding year. source
Customer-controlled storage
For production, documentation recommends retaining data in the customer's own bucket and sending URLs, including presigned or cross-account URLs. source
Documentation coverage
The documentation includes getting-started tutorials, API reference, API-key setup, import guidance, cloud integrations and Python SDK instructions. source
Enterprise access
Enterprise includes SSO and MFA, custom security solutions, and a dedicated solutions ML engineer. source
Founded
2020 source
Founders
Y Combinator identifies Otto Debals and Bert De Brabandere as founders; Crunchbase also lists Davy Neven among the founders. source
Funding
Y Combinator says the company was funded in 2021 by Y Combinator, Merus Capital and Volta Ventures. source
GetApp listing
GetApp's October 2026 listing shows no user reviews and a 0.0 score for ease of use, features, value and support. source
Integrations
The Python SDK supports programmatic data upload, label downloads, dataset management, and webhooks; the integrations page names TensorFlow and PyTorch. source
Intended users
The company describes the product as designed for computer vision engineers and data scientists working in autonomous robotics and driving. source
Legal entity
Crunchbase identifies the legal name as Segments.ai BV and lists Leuven, Belgium as headquarters. source
Limit: heavy 3D workloads
Individual G2 reviews report slow or stuck 3D visualization on computationally heavy or large point-cloud sequences. source
Limit: learning curve
G2 review summaries repeatedly mention a steep learning curve for newcomers, complex features and insufficient guidance. source
Limit: training and exports
G2 reviewers report limited tutorials, a lack of training material and a desire for better direct web export options. source
ML-assisted labeling
Image tools include Superpixel, Autosegment, hover-and-click segmentation, and uploaded model predictions for review and correction. source

Checked against the maker's pages on 15 Oct 2026.