From application to payout
Seven stages, each one something the platform genuinely does today. No stage below is aspirational and none of it is outsourced to an algorithm making decisions about you.
Apply to a listing
Create an account with your country of residence and nationality. Those two fields decide which work you can legally be offered, so listings you aren't eligible for never appear at all. You can't waste an application on one.
You can upload a CV and have the structured fields pulled out for you or fill your profile in by hand. Either way the profile is built once and reused for every listing you apply to afterwards.
Sit a structured interview
A short set of written prompts, scored against a rubric written by Corpshore AI staff. Scoring assists the decision. It never makes it.
A weak score routes your application to a human for a closer look rather than auto-rejecting you. Brief section 0.9 is explicit about this and the platform enforces it: automation may only hold, flag or reject nothing at all. Humans approve.
Pass the project's course
Many projects include a short course covering that project's exact conventions. Tagging schemes, capture checklists, rubric definitions, with a gating assessment at the end.
Passing enrols you in the project automatically. What you're taught in the course is precisely what QA reviews against, so there's no hidden standard waiting to catch you out later.
Reserve work when you want it
Open tasks appear in your dashboard with their pay rate showing. Reserve one and it's held for you. Nobody else can take it while you're working on it.
Reservations carry a time window so queues keep moving, but there are no shifts and no minimum hours. Work at 2am on a Sunday if that's what suits you.
Submit, with the tools doing their share
Capture tasks validate photos as you shoot them. Transcription tasks can generate an AI first draft for you to correct. Evaluation tasks show the frozen rubric next to the content.
Validators catch problems before submission rather than after rejection. AI assistance is always a draft you review and edit. It is never auto-submitted on your behalf, on any task type.
Human quality assurance
A QA reviewer checks your work against the project standard. Accept, reject with a written reason or request a cure.
A rejection cites the specific criterion and opens a cure window so you can correct and resubmit. If you believe the reviewer got it wrong, a team lead can overrule the decision and that overrule is logged.
Get paid
Acceptance credits your ledger immediately. Payouts run in a weekly batch through the rail available in your country.
You're paid on our internal QA acceptance, never on a client's review timeline. So a slow client can't delay your money. Payout batches require dual approval from two different administrators before release.
That's the whole loop
Apply once, qualify once, then work as much or as little as you want.