An enterprise data-labeling SaaS built by TensorMatics for teams training computer vision and NLP models. It supports images, video, PDFs, text, and audio, with annotation types including bounding boxes and keypoints. Its core pitch is that you annotate only 10–20% of a dataset and let prompt-based labeling, model-assisted annotation, and active learning complete the rest. It integrates with GCP Vertex AI, AWS SageMaker, and Azure ML, and exports to CSV, JSON, COCO, Pascal VOC, or custom formats. Security includes AES-256 encryption, TLS 1.2+, Auth0 authentication, and optional customer-owned cloud hosting. Pricing data conflicts sharply across sources—Capterra lists ₹1,000/user/month, SaaSworthy lists Starter $49/month, Crozdesk lists StartUp $600/month, and Toolradar lists Pro at $9,999/year—so the official pricing page should be the reference. Company fundamentals are also inconsistent: founding year is listed as 2017, 2018, 2019, or 2020 depending on the source, and headquarters appear as either San Francisco or Mohali, India. PitchBook shows roughly $166K raised across accelerator and incubator rounds, with a team of 13–16 people. Indian Express reported in September 2025 that revenue was on track to exceed $1M that year. Capterra shows 17 reviews and G2 shows 21 reviews at 4.8/5, though many reviewers are India-based and some reviews date to 2021–2023.
**Pros & Cons**:
Pros:
ML-assisted labeling with active learning aims to cut annotation volume to 10–20% of the dataset, which is the main cost driver in labeling projects; supports images, video, PDFs, text, and audio in one platform rather than requiring separate tools per modality; MLOps integrations with Vertex AI, SageMaker, and Azure ML fit existing training pipelines; export formats cover COCO, Pascal VOC, CSV, JSON, and custom schemas; AES-256 encryption, TLS 1.2+, and optional customer-owned cloud hosting address data-handling requirements; a free Researcher tier exists, though its allowance varies by source.
Cons: **Pricing information conflicts severely across sources**—₹1,000/user/month, $49/month, $600/month, and roughly $833/month annual all appear, so no single figure should be treated as accurate without checking the official page; company fundamentals are inconsistent—founding year spans 2017–2020 and headquarters appear as both San Francisco and Mohali, India, which complicates verification; funding is small (~$166K) with a 13–16 person team**, so long-term scale is unproven despite reported revenue growth; **review samples skew India-based and partly outdated**, with some Capterra reviews dating to 2021–2023; users have reported delays between submission and frame loading, a missing undo button, and early-product slowness that improved over time; free-tier allowances differ by source (1,000 units/month vs 2,500 files), so the official page should be checked.
**Who is it for?**:Computer vision and NLP teams that need to annotate large datasets without building an in-house labeling tool; ML teams already using Vertex AI, SageMaker, or Azure ML that want annotation integrated into the training pipeline; organizations that require encrypted storage and optional customer-owned cloud hosting; small teams that want a free Researcher tier to evaluate before committing.