By Mufy Pachorawala, founder of getAIwork · Last updated: August 27, 2026
Key facts
- Both use AI-driven interviews: Mercor’s roughly 20-minute video interview with retakes; Micro1’s adaptive interviewer “Zara” with integrity monitoring
- Neither charges workers any fee; both take their margin from the client side of each placement
- Scale in mid-2026: Mercor reported a $2B gross annualised run rate (July); Micro1 announced a $500M gross run rate (August)
- Payment terms differ concretely: Mercor weekly via Stripe or Wise; Micro1 twice monthly in USD via Deel with KYC and a background check first
- 630 platform programmes are live on the getAIwork board as of August 22, 2026, screened and human-approved
What do Mercor and Micro1 have in common?
They are the two defining names of the expert tier of AI training work, both built on the same 2025 shift: AI labs stopped buying mass generalist annotation and started buying verified expertise. Both companies recruit globally, screen with AI rather than human recruiters, place accepted experts into lab projects, pay real money reliably, and reject or shelve most applicants without much communication. Both also monetise identically from a worker’s point of view: you pay nothing at any stage, and each company takes its margin from what the client pays. If you remember one shared fact, make it that one, because every site charging an application fee in their name is an impersonation.
How do the applications differ?
Mercor: profile plus resume, then an automated video interview of about 20 minutes with retakes allowed, then algorithmic matching against open projects. The interview probes domain depth conversationally, and there is no live coding requirement for non-engineering roles.
Micro1: resume plus skills profile, then an interview with Zara, its AI interviewer. For engineering roles this is the tougher screen in the pair: roughly 20 to 30 minutes, adaptive, often with live coding, and monitored by an integrity system that tracks tab switches and attention. Accepted experts then apply to specific projects rather than waiting for pure matching, and complete identity verification and a background check before first payment.
Net: Mercor’s screen is friendlier for non-technical professionals; Micro1’s is built to certify engineers, and its anti-cheat layer makes shortcuts a losing strategy.
Answer 6 questions and see which of the 630 live platform programmes on the board actually match your profile before you spend interview effort.
Take the 2-minute match quiz →
How do pay and payment terms compare?
| Dimension | Mercor | Micro1 |
|---|---|---|
| Listed or reported rates (2026) | $63 to $120/hr listed roles; reported spread roughly $20 to $200+ by domain | $40 to $100+/hr reported for engineering and AI training; up to $500/hr reported for rare medical and finance specialisms |
| Payment cadence | Weekly (Sat to Fri week, paid around Wednesday) | Twice per month, in USD |
| Payment rails | Stripe; Wise where Stripe unavailable; local currencies | Deel, after KYC and background check |
| Worker fees | None, by published policy | None reported anywhere |
| Time tracking | Desktop app with periodic screenshots on hourly work | Per-project hourly, per-task or flat-fee models |
| Reviews (Aug 2026) | Trustpilot 4.1/5, about 226 reviews | Trustpilot around 4.7/5 (it sat near 2.2 in October 2025); Glassdoor 4.5/5 |
Read the rates the way this whole category demands: figures are per project, as listed or as reported at specific dates, and none is guaranteed. Micro1’s review-score swing is worth naming rather than hiding: the climb from roughly 2.2 stars in October 2025 to about 4.7 in August 2026 coincides with its contractor base growing rapidly, and whichever explanation you prefer, both dates are real data.
Which platform has more work right now?
Mercor is the bigger machine in August 2026: a reported $2 billion gross annualised run rate against Micro1’s $500 million, tens of thousands of weekly active experts, and a visibly wider spread of non-engineering roles (medicine, law, finance, marketing, even architecture). Micro1’s demand concentrates where its engineering-first pipeline points: code evaluation, agent training, and its newer robotics data programmes. Neither publishes queue depth, and on both the accepted-but-idle experience is the top complaint, so the practical read is: more professional variety at Mercor, more engineering certainty at Micro1, and no guarantee of steady volume at either.
Company size does not pay your invoice; live queues do. See which programmes currently have work matching your skills.
Find your match in 2 minutes →
Which should you choose?
| Your profile | Better first bet | Why |
|---|---|---|
| Licensed professional (medicine, law, finance, accounting) | Mercor | Its listed role board is built around professional studies at $63 to $120/hr |
| Software engineer or ML practitioner | Micro1 | Zara’s technical screen is the product; engineering demand is its core |
| Strong generalist, no credentials yet | Neither first | Both will likely accept and shelve you; earn on open-intake platforms while building a specialism |
| Academic or researcher | Mercor, then Micro1 | Research studies appear on both; Mercor’s variety gives more surface area |
| Non-US applicant | Either | Both recruit globally; check each project’s country and language requirements |
The apply-to-both playbook
The interviews are short, free and non-exclusive, so for anyone with a real specialism the optimal play is both, deliberately: take Mercor’s interview in a quiet hour with concrete case examples ready; take Zara after warming up on the kind of live problem your domain implies; then stop watching either inbox. Silence is the default state on both platforms and says nothing about you. Keep an open-intake platform paying in the meantime, per our no-experience guide, and let whichever expert queue activates first claim your hours. Full platform context sits in our 12-platform comparison and Mercor review.
Mercor, Micro1 and 630 other live programmes are tracked on the board, with intake status checked daily and filled listings deleted.
Frequently asked questions
Are Mercor and Micro1 both legit?
Yes. Both are heavily funded AI data companies (Mercor valued at $10B in October 2025; Micro1 at $500M in September 2025) with consistent worker reports of real payouts, and neither charges workers anything. The shared complaint is silence after screening, not non-payment.
Do Mercor or Micro1 charge any fees?
No. Mercor’s published policy is that it never requires payment from talent, and no credible report shows Micro1 charging workers either. Both earn from the client side. Treat any fee request in their names as an impersonation scam.
Which pays more, Mercor or Micro1?
Overlapping bands with different peaks: Mercor lists roles at $63 to $120 per hour across professions, while Micro1 reports $40 to $100+ for engineering with rare specialisms reported far higher. Your domain decides which ceiling applies to you; neither guarantees volume.
What is Zara on Micro1?
Micro1’s AI interviewer: an adaptive, recorded interview, often with live coding for technical roles, paired with an integrity system that monitors attention and tab switching. It is the main gate into Micro1’s expert network, so treat it as a full technical interview.
Can I retake the interviews?
Mercor allows interview retakes. Micro1 retake policies vary by role and project; a failed technical screen generally means waiting before another attempt, so warm up first rather than treating the real interview as practice.
Which platform responds faster after the interview?
Both are demand-driven and neither promises a timeline. Reports cluster around two to four weeks for successful Mercor matches and similar or longer for Micro1 project starts, with many accepted applicants on both hearing nothing for months. Parallel income elsewhere is the sane hedge.
I am not an engineer and have no license. Should I apply anyway?
The interviews cost only time, so you can, but set expectations: both platforms shelve most generalists. The faster route is earning on open-intake platforms now and building documented specialist evidence, then returning to Mercor and Micro1 with something their matching engines can price.
Mufy Pachorawala · Founder, getAIwork
My AI scans thousands of AI-job posts (38,434 screened so far) and I personally approve every listing before it reaches the board. I write these articles by the same rules: pay quoted only as listed, and no income promises. Read our editorial rules.
Related: Mercor review · 12 Outlier alternatives compared · AI training jobs with no experience · Outlier AI review
