Labelbox

Labelbox is a data platform for AI that pairs a multi-modal data-labeling and annotation platform with on-demand human labeling services, used to create high-quality training and evaluation data for ML and generative AI. It supports images, video, text, and documents, and adds workflow, quality control, model-assisted labeling, and data curation. Used by large enterprises, it has raised roughly $189M total, including a $110M Series D led by SoftBank Vision Fund 2.

Labelbox helps teams build the labeled datasets that AI models depend on. Its platform provides annotation tools for many data types (images, video, text, documents, and more), workflow and quality-control features, model-assisted labeling, and data curation to prioritize the most valuable examples. Over time it has expanded from classic supervised-learning labeling toward generative-AI data needs like human preference data, RLHF-style feedback, and evaluation. A distinguishing element is that Labelbox pairs its software with an on-demand labeling workforce ('Boost'/human-data services), so customers can either label in-house using the platform or outsource labeling to Labelbox-managed labelers. This 'data factory' positioning targets enterprises that need large volumes of high-quality labeled or human-feedback data without building their own annotation operations. Labelbox is used by large enterprises across industries and reports meaningful revenue scale. It has raised a total of roughly $189 million across several rounds, with a $110 million Series D led by SoftBank Vision Fund 2 (2022) and backing from investors including Andreessen Horowitz, B Capital, Gradient Ventures, and Databricks Ventures. Pricing blends a self-serve, pay-per-usage model (priced in 'Labelbox Units') with sales-led enterprise contracts and separately quoted human-data services.