We let professionals with medical judgement take part in medical data annotation and earn rewards in their own fragmented time. The platform standardises annotation work and makes it payable per piece, and provides identity verification, online annotation tools, multi-stage quality control and settlement guarantees —— participants' professional judgement is priced accurately and data quality is strictly safeguarded.
A tiny lesion in an image, an abnormal interval in a waveform, a key sentence in a medical record —— these judgements require long professional training. The Crowd Platform breaks this work into standardised, verifiable task units, so physicians, medical students and professional annotators can take part on their own schedule and earn rewards that match their expertise.
Each task's annotation unit price, estimated hours, acceptance criteria and QC standards are fully displayed before claiming. Once completed, pricing follows actual quality, and reward details can be checked item by item.
Tasks are released in batches and can be claimed at any time and completed in stages. Most single tasks take 10–40 minutes — suited to gaps between clinics, downtime on shift and fragmented evening hours.
From basic annotation to specialist annotation and then quality-control review, your level is decided by actual annotation quality and pass rate; the higher the level, the wider the range of tasks you can take on and the higher the unit price.
Medical data annotation is special in that a single mistake can affect a model's judgement. The platform therefore sets hard constraints at three levels.
After submission a machine pre-check runs first, then cross review; high-risk tasks receive final review by a specialist physician, and only annotations that pass count towards reward.
Annotation takes place in a controlled environment, task data is displayed masked, and participants cannot export, download or retain any raw medical data.
Every claim, annotation, revision and review leaves a timestamp and operation log, forming a traceable chain of evidence.
Different professional backgrounds map to different task levels. You can stick to basic annotation, or take on specialist annotation and quality-control review once you pass the assessment.
Take on specialist annotation and final review of imaging, waveforms and medical record text. Suited to radiology, ultrasound, obstetrics and gynaecology, urology, rehabilitation and similar fields, and can be done in gaps between clinics and on shift.
Advanced tasks · highest unit priceComplete basic and advanced annotation tasks and learn by doing: the platform provides annotation specifications and an example library, and passing the trial annotation assessment lets you take on imaging outlining, text structuring and similar tasks.
Beginner-friendly · specifications you can learnThose with annotation industry experience can take the advancement assessment directly and take on large-volume, clearly specified tasks; long-term stable participants can apply to become dedicated batch contractors.
Batch tasks · steady supplySelected from participants with a consistently high pass rate, they handle cross review and dispute decisions, and are paid by review volume with a quality bonus.
Requires a qualifying quality recordFirst complete identity verification to gain access, then enter the online annotation workspace to claim tasks. Every step has clear rules and visible progress.
Access is granted in four steps: real-name verification, practising and qualification checks, a trial annotation skills assessment and signing a compliance commitment. Passing gives you a level that matches your ability, and the level decides which tasks you can take on and their unit prices.
Complete all work in the browser: filter and claim in the Task Hall, online annotation with keyboard shortcuts, automatic pre-check, submit for review, view reward details and level progress —— no client installation needed.
Every step has a clear time expectation and status indication — you will never be left wondering what comes next after submitting.
Register with your mobile number, complete real-name verification and set up a settlement account. About 10 minutes.
Submit practising or academic and professional credentials, complete specification training and the trial annotation assessment. Feedback within 1–2 working days.
Filter by modality, difficulty and unit price in the Task Hall, check the pricing and acceptance criteria, then claim.
Complete annotation in the workspace; the system pre-checks in real time and flags suspected errors or omissions, which you can correct before submitting.
Once review is passed the work counts as valid workload and enters settlement on quality-based pricing; reward details can be checked item by item.
The task type decides the background required and how it is priced. You can focus on just one type, or expand gradually as your level rises.
Outline, measure and grade target structures in ultrasound, X-ray, CT, MRI and other imaging. These tasks demand the highest anatomical and image-reading ability, and carry the highest unit prices.
Annotate time intervals and event types on synchronously acquired physiological waveforms, such as parameter-change segments during treatment, or event points on pelvic floor EMG and pressure curves.
The workspace has built-in waveform zoom, rulers and shortcut annotation, supports millisecond positioning, and annotations can be reviewed and revised at any time.
Extract structured fields from masked clinical course records and examination reports, annotating entities, relations and negation / suspicion semantics.
Text tasks are characterised by clear rules and ease of entry; they are usually the first task type for new participants and the main route to raising your level.
Transcribe follow-up calls and consultation recordings, separate speakers and extract key information, to build follow-up datasets and adherence analyses.
Tasks involving speech require an additional confidentiality and privacy commitment; the speech provided by the platform has already been masked and stripped of identity information at the point of capture.
This is both a quality requirement and a protection for participants: clear acceptance criteria mean no wasted rework because subjective standards were unclear.
On submission the system validates automatically against the rules: field completeness, reasonableness of the outlining range, time-point boundaries and logical consistency; suspected errors or omissions are flagged for you to correct.
Blind review is carried out by participants at the same or a higher level; agreement means it passes. Any disagreement goes into the dispute pool for a higher-level reviewer to decide.
High-risk and specialist tasks undergo final review by clinicians, with a proportion sampled for full re-check, forming the final quality-control conclusion and the basis for pricing.
Primarily per-piece, weighted by quality and settled on schedule. Unit prices vary with difficulty and the professional level required, with no hidden deductions.
* Actual per-task unit prices, settlement cycles and tax handling are subject to the platform's task descriptions and the service agreement signed by both parties. Income earned by participants on the platform is personal service income, and related taxes and fees are handled in accordance with the law.
Yes. The platform provides annotation specifications, an example library and operational training, and completing the trial annotation assessment qualifies you for basic tasks. Medical students and residents usually start with text structuring and basic outlining tasks, and level up gradually once their pass rate is stable.
There is no required online time and no daily minimum task volume. Tasks are released in batches; you can claim one at a time, or take a batch and finish it over several days. Unfinished tasks are automatically reclaimed and re-released.
No. Data in tasks is masked on the principle of minimum necessity and displayed in a controlled environment: only the content needed for annotation is shown, with no names, ID numbers, contact details or other personally identifiable information, and exporting or retaining it is not permitted.
Tasks are priced per item; once review is passed they count as valid workload and are settled in aggregate at the agreed cycle. Reward details (task ID, unit price, status, review conclusion) can be checked item by item in the workspace, and any discrepancy can be appealed.
Participants take on platform tasks as individuals; they are not tied to the data ownership of any specific care provider, nor do they bear the responsibilities of a data holder. What circulates on the platform are processing tasks that have been masked and authorised. If your institution has separate rules on outside part-time work, please confirm and comply with your employer's requirements first.
Whether you are a clinician, a medical student or an experienced professional annotator, completing identity verification lets you view the currently open tasks and pricing standards.
