Label Less is a Hong Kong social enterprise focused on data annotation and human review, with a mission to open up opportunities through human work in AI.
Label Less’s completed pilot was part of MakerBay’s Impact Innovation Lab, Cohort 4, with support from the SIE Fund. This acknowledgement refers to that pilot.
Free practical training starts from the workflows your team already handles: forms, attendance and activity records, grant reports, receipt checks and internal tools. We plan each workshop and date individually.
Explore NGO AI workflow training →Download the sample CSV and bilingual review checklist. Practise checking duplicates and missing attendance statuses before drafting a report.
Explore the attendance example →Label Less is a Hong Kong social enterprise focused on data annotation and human review. We aim to open up opportunities through human work in AI.
Teams exploring data annotation for AI and research workflows.
Label Less’s core work is data annotation and human review.
→ →Label Less is a social enterprise, not an NGO. Its core work is data annotation and human review.
Bounding boxes, polygons, semantic segmentation, keypoints and classification for computer vision.
Cantonese and Traditional Chinese, Hong Kong nuance, NER, sentiment, intent and content moderation.
Transcription, speaker labeling, timestamping and acoustic event tagging across dialects.
Sensor and LiDAR labeling, trajectories, action segments and teleoperation demonstrations.
Wearable-camera capture of first-person task demonstrations, hand-object interactions and trajectories for embodied AI and robot foundation models.
Review steps and quality checks are agreed for each proposed annotation task.
We select for care, focus and commitment. Those traits turn straight into accurate labels.
For each proposed project, the review approach and data handling are discussed before work is scoped.
Annotators learn your spec and edge cases.
Careful labeling at a pace that protects quality.
Multi-pass QA and consensus before sign-off.
Model-ready data in your format, on schedule.
Our history
MakerBay’s August 2025 account of Label Less’s Cohort 4 pilot reported that about 30 mothers were recruited for data-labelling training. That programme has ended; this is historical context, not a current placement or partnership offer.
Read MakerBay’s historical accountThe name is also a promise: to put fewer labels on people. Single parent. Disabled. Patient. We help shed those labels by labeling the data that powers AI.
Label Less is a Hong Kong social enterprise. Data annotation and human review are its core work, and its mission is to open up opportunities through human work in AI.
Read the historical account ↗Tell us the data type, format and scope you have in mind. We will discuss feasibility and the next step.
Request a quote →Contact us to discuss Label Less’s work and social enterprise mission.
Get involved →Data annotation, or data labeling, is the process of tagging images, text, audio, sensor and robotics data so machine-learning models can learn from them. It is the human foundation that makes AI accurate, and it is exactly what we do.
Image and video, text and language with Cantonese and Traditional Chinese local context, audio and speech, robotics and teleoperation data such as sensor, LiDAR, trajectories and action segments, and egocentric, wearable-camera capture of first-person task demonstrations.
For a proposed annotation task, we can discuss written guidelines, review checkpoints and how quality would be checked before any work is scoped.
Label Less is a Hong Kong data annotation and human-review social enterprise. Staffing and any community support for a future project would be discussed as part of its scope.
Plan data handling for each workflow. Review data minimisation, processing locations, retention and training terms, permissions, source traceability and staff review before connecting tools. No workflow is risk-free or a compliance certification.
Tell us the data type, format and scope you have in mind. We will discuss what may be feasible; availability and timing are confirmed individually.