Data Entry Keyer

$20–35/hrrate as listed

Transcribe and structure source material — scanned documents, forms, tables, handwritten notes, screenshots — into clean, schema-conformant records used to train and evaluate AI systems. Work is task-based and remote, with observed rates in the $20–35/hr range depending on batch complexity and accuracy history.

Summary, eligibility guidance and screening preparation are prepared by Sidequest from the platform's posting. Check the source listing before applying.

What the work involves

You receive batches of source material and a field spec, then key the contents into a structured form or spreadsheet-style interface. Sources vary: scanned invoices and receipts, government or medical forms, historical records, handwritten survey responses, chart images, or screenshots of software UIs. Some batches ask for verbatim transcription; others ask you to normalize (dates to ISO format, currencies to a standard notation, names to last-first) according to written guidelines.

A meaningful share of the work is second-pass review rather than first-pass keying — comparing an existing record against the source and flagging mismatches, or adjudicating where two earlier keyers disagreed. You will also hit ambiguous cases: illegible handwriting, fields cut off by the scan, values that contradict each other within the same document. The expectation is that you follow the escalation path in the guidelines rather than guessing, and that you can explain your reasoning in a short note.

What the platform screens for

Sustained accuracy over speed. Screens typically include a timed keying sample scored on character-level and field-level error rate. Consistency across a long batch matters more than a fast first minute.
Guideline literalism. Can you read a field spec and apply it exactly, including the edge cases it names — and recognize when a case falls outside it?
Legible judgment on ambiguity. Interviews probe how you handle unreadable or contradictory source data, and whether you flag rather than fabricate.
Tooling comfort. Keyboard-first workflows, spreadsheets, keyboard shortcuts, and occasionally light regex or find-replace for cleanup.
Confidentiality discipline. Batches may contain PII; expect NDA terms and rules against local copies or screenshots.

Logistics

Fully remote and largely asynchronous. Work is released in batches with deadlines rather than fixed shifts, so hours are flexible, though some clients cluster releases in US business hours and batches can expire if unclaimed. Most contributors work 10–25 hours per week; volume is genuinely variable and there is no guaranteed weekly minimum. Rates observed in this band vary with document difficulty, language, and your accumulated quality score — higher rates generally attach to specialized sources (medical, legal, non-Latin script) or to review and adjudication tiers.

What they look for

Accuracy discipline
Error detection
Format reconciliation
Attention to Detail

Before you apply

Would you pass this screen?

Drawn from what this listing actually requires. Answered in your browser — nothing is sent to us, and we do not store your answers.

Are you comfortable typing continuously for 2–3 hour stretches and proficient enough with keyboard shortcuts and spreadsheet software that you don't need tutorial support to begin?
Keyboard speed and comfort · preferred
Do you have a reliable internet connection and a private, quiet workspace where you can work on confidential documents without risk of observation or local backup?
Remote work setup · required
Can you read and follow written instructions in English, including recognizing ambiguities and edge cases in field specifications?
English reading comprehension · required
Can you claim and complete work batches on flexible schedules with variable deadlines, rather than fixed shifts, and commit to submitted deadlines?
Asynchronous availability · preferred

0 of 4 answered

Included with this listing

Interview prep — 10 questions

Matched to what this platform's screening actually measures. Answer each out loud in 60–90 seconds: claim → concrete example → implication.

What it tests: Whether you follow the escalation protocol (flag and document) rather than guess, and whether you can articulate reasoning that respects both data integrity and the guidelines you're given.
A strong answer: You'd note exactly what's illegible, explain why you can't infer the month safely, and follow the guideline's escalation step—either flagging for supervisor review, leaving it blank per spec, or marking it as partial/uncertain. The key is showing the work, not filling a gap.
Where people lose it: Saying you'd 'just take your best guess' or 'fill in what makes sense' signals you prioritize speed or completeness over accuracy and guideline fidelity—the opposite of what the role demands.