At-Home Video Recorder

$13–15/hrrate as listed

Record short at-home videos to spec — everyday tasks, hand movements, object interactions, or scripted scenarios — that become training and evaluation data for multimodal AI models. Work is asynchronous and paid hourly at an observed $13–15/hr on micro1, with volume depending on active data-collection campaigns.

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

Multimodal models need footage that public video datasets don't cover: a person's hands assembling something on a cluttered kitchen counter, a household object filmed from six angles under uneven light, a task performed correctly and then performed with a deliberate mistake. This role is the supply side of that. You receive a capture brief — subject, camera framing, duration, lighting conditions, number of takes, what must and must not appear in frame — and you produce clips that satisfy it exactly, then upload them with the metadata the brief asks for.

Day to day, that looks like: reading a spec, setting up a phone or webcam on a tripod or improvised mount, doing several takes, reviewing your own footage for blur, clipping, audio problems, or accidental PII (mail, screens, faces of bystanders), naming and tagging files per convention, and submitting. Some campaigns are single-shot and repetitive; others are scripted or multi-step and ask you to narrate what you're doing. Rejected clips usually come back for a specific, fixable reason — wrong aspect ratio, hands leaving frame, background too dark — so the people who do well here are the ones who read the brief closely the first time.

What the platform screens for

micro1's screen is conversational and AI-led. Expect it to check that you actually have a usable capture setup (device, resolution, stable mount, a space you can film in without other people wandering through), that you understand instructions literally rather than approximately, and that you'll flag an ambiguous brief instead of guessing. Reliability and turnaround matter more than production polish — nobody is grading your cinematography. Consent and privacy awareness comes up: you must be the only identifiable person in frame unless the brief says otherwise, and you need to be comfortable with your footage being used as AI training data.

Logistics

Fully remote, filmed wherever you live; no scheduled calls once onboarded.
Asynchronous, with per-batch deadlines typically measured in days rather than hours.
Volume is campaign-driven — some weeks are busy, others quiet. Treat it as supplemental rather than a fixed schedule.
Pay observed at $13–15/hr, not guaranteed; some campaigns are structured per approved clip instead of hourly.
You supply your own phone or camera, storage, and internet bandwidth for uploads.

What they look for

Smartphone
Video Recording
Attention to Detail
Household Tasks

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.

Do you have access to a smartphone or camera capable of recording at least 1080p video resolution?
Video capture device · required
Can you stably mount your recording device (phone or camera) on a tripod, stand, or other fixed support without holding it in your hand during recording?
Stable mounting · required
Do you have a space in your home where you can film alone without other identifiable people walking through during daytime hours on most days?
Solo filming space · required
Do you have internet access capable of reliably uploading video files of several hundred megabytes in size?
Internet upload capability · required
Do you understand and consent that video footage you record may be used as training data for AI models and will be shared beyond the micro1 platform?
Privacy understanding · required

0 of 5 answered

Included with this listing

Interview prep — 11 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 the candidate treats ambiguity as a blocker that requires clarification rather than a problem they'll solve by guessing or over-interpreting the intent.
A strong answer: A strong answer names the specific ambiguity, explains why it matters to the end result, and states they would ask the requester to clarify before filming — not make an assumption about what 'counts' as properly done.
Where people lose it: Candidates who say they'd film both versions, or film one version and hope it's right, miss the point: the role is about exact compliance, not hedging bets or reading minds.