Encord iconEncord

Encord is a data development platform for physical and multimodal AI: curation, annotation and model evaluation for video, LiDAR, DICOM and more

$60M Series C (Feb 2026)
300+ Physical AI Teams (company claim)
4.8/5 G2 Rating

Overview

Encord is a data development platform that prepares training data for AI, with a strong focus on physical AI such as robots, drones and autonomous vehicles. It brings data management, curation, annotation and model evaluation into one platform, organized as Index (curate and filter raw data), Annotate (create labels) and Active (analyze label and model quality at scale), with an SDK for keys, data registration and label import and export. It handles an unusually wide range of data: images, image groups and sequences, video, audio, PDFs, text, DICOM and NIfTI medical images rendered natively in the browser, point clouds (PCD, PLY, LAS, E57, ROS bag, MCAP in beta) and time series. In February 2026 Encord (legal name Cord Technologies Inc.) raised a $60M Series C led by Wellington Management, bringing total funding to $110M; Y Combinator, CRV and N47 are earlier backers. Co-founders and co-CEOs are Eric Landau and Ulrik Stig Hansen. The company says data on the platform grew from just over 1 PB to more than 5 PB and revenue grew more than 10x in about 18 months (company claims).

The Verdict

Who Should Use Encord?

Best For

  • Robotics, drone and autonomous vehicle teams with camera, LiDAR and sensor data
  • Medical imaging teams that need native DICOM and NIfTI annotation and review trails
  • Video-heavy computer vision projects

Not Ideal For

  • Teams that want an open-source, self-hosted tool (see Label Studio)
  • Buyers who need a public price list before talking to sales
  • Pure text or LLM preference work, where lighter tools exist

What's Great

  • Support, ease of use and task quality score highly on G2
  • Video-first annotation and multi-stage peer review
  • One platform for curation, labeling and model evaluation
  • Native point cloud, DICOM and NIfTI support

Watch Out For

  • Pricing is the most common complaint on G2
  • Some reviewers find automation complex and hotkey customization limited
  • No public pricing; G2 lists the entry price as "Contact Us"

Pricing

View all features & details

Platform

  • Index: organize, filter and curate raw data
  • Annotate: labeling with review workflows
  • Active: label and model quality analysis
  • SDK: access keys, cloud data registration, label import and export
  • 2026 updates: Labels in Index (May 2026), audio editor and custom annotation layouts

Supported Data

  • Images, image groups, image sequences
  • Video (MP4, WebM, MKV) and audio (WAV, FLAC, MPEG)
  • PDF documents and text (HTML, JSON, XML, TXT)
  • DICOM and NIfTI
  • Point clouds (PCD, PLY, LAS, LAZ, E57, ROS bag, MCAP beta) and time series CSV

Real-World Usage

Named Customers

  • Woven by Toyota
  • Zipline
  • Skydio

Scale and Reviews

  • Works with more than 300 physical AI teams (company claim)
  • G2: about 4.8 out of 5 from about 65 reviews (mid-2026)

How It Compares

Feature Encord Labelbox Label Studio Snorkel AI
Focus Physical AI and multimodal data platform Frontier data, RL environments, labeling Open-source labeling and AI evaluation Expert data-as-a-service
Open source No No Yes, Apache-2.0 Old snorkel library only
Free option Not confirmed 500 LBU a month Community Edition No
Public pricing No In docs Yes No
G2 rating About 4.8 About 4.5 Too few reviews Too few reviews

Sources: Encord docs · Labelbox limits · HumanSignal pricing

Is this your tool? Add a badge