Platform review
micro1
Pass a live AI interview, then wait to be matched. Published bands run $15 to $200+ an hour.
micro1 is a credential-first platform. You sit a live AI interview before you see any work, and the money sits in the upper bands where a real domain credential is the entry ticket. It publishes pay bands by task type, from $15 to $40 an hour for general annotation up to $100 to $200+ for frontier expert projects, which is more disclosure than most of this category offers. What it does not offer is a start date: certification makes you eligible and nothing more, either side can end the contract on five business days' notice without cause, and a pipeline runs only as long as the client has work. Worth the 20 to 40 minutes if your credential reaches the top two bands, not if you are aiming at the floor.
Median listed rate
$70/hr
The median listing's own midpoint, across the 313 of 329 open roles that list an hourly rate. Half sit above it, half below; the middle half run $40–$105/hr and the full spread is $6–$400/hr. This is what micro1 advertises, not what anyone has been shown to earn.
Application style
AI interview
micro1's own reviews page publishes expert testimonials describing an AI-conducted interview step. How hard it is to pass, we can't tell you yet.
Specialties on its board
13 sections
Computed from micro1's own specialty grouping across its 329 open roles — only sections with enough priced listings to say anything about are counted.
Key facts about micro1
- Pay
- $70/hr median ($40–$105 middle half)
- Application
- Interview / vetted (per platform)
- Availability
- Global
- Specialties
Read this first
The things that change how you should read everything else on this page.
- The 4.7 Trustpilot score is a rating of micro1's AI interview, not of working for micro1. Wherever that number appears, this caveat has to travel with it.
- micro1's expert agreement requires you to warrant that you are not currently employed by a foundational AI model company. If you work at one, you are excluded, and the marketing pages do not say so.
- micro1's onboarding page lists Karma Check, a third-party background check, as part of onboarding. It publishes no detail on what is checked or who sees the result.
- micro1's own documents contradict each other on pay frequency: the experts guide says biweekly, the payroll policy says semi-monthly.
- Hourly pipelines require the Hubstaff desktop tracker, and micro1 publishes no policy on screenshots, activity monitoring or idle time.
- Trustpilot reviewers in the 1- and 2-star band repeatedly allege that the interview is used to collect expertise rather than to hire. That is an allegation by reviewers, not a finding, but it recurs often enough to be worth knowing before you spend 40 minutes answering technical questions out loud.
What the work is
micro1 sells human data and evaluation products to AI labs, marketed as Realm (reinforcement-learning environments), Cortex (contextual evaluation) and a robotics data line. The worker-facing half of that business is an expert network: people with subject-matter credentials produce and check the data those products are built from, engaged as independent contractors on client projects micro1 calls pipelines. Its worker documentation splits the available work three ways: frontier evals, which are expert-led evaluations used to train foundational models; real-world robotics, where contributors record everyday physical tasks; and contextual evals built around a client's own workflows.
Who they want
micro1 recruits domain experts across engineering, finance, healthcare and medical, legal, economics, linguistics and cinematography, and says its opportunities run from early-career professionals through PhDs and industry leaders. Its own pay guide implies the same spread, with general annotation at the bottom and frontier expert projects at the top. The expert agreement adds one exclusion the marketing pages leave out: you warrant that you are not currently employed by a foundational AI model company.
Step 1
Apply in the Opportunities Portal
Each role carries skill requirements micro1's recruiters set from the client's brief.
Step 2
Live interview with Zara
micro1's AI recruiter, spoken and open-ended: 20 to 40 minutes, about seven per skill.
Step 3
Certification
Puts you in the certified talent pool, which micro1 says does not guarantee placement.
Step 4
Matching to a pipeline
By invitation as relevant roles open, or by applying directly to portal postings.
Step 5
Onboarding, then a first task
Slack, an expert email, Hubstaff time tracking and a Karma Check background check.
How you get in
Applications run through micro1's Opportunities Portal, where each role carries skill requirements its recruiters set from the client's brief. Everyone who applies is interviewed by Zara, micro1's AI recruiter, in a live session where you answer open-ended questions out loud. It runs 20 to 40 minutes depending on how many skills you are being assessed on, at roughly seven minutes per skill, and micro1 offers a free 18-minute practice interview tailored to a chosen role if you want a dry run first. Passing puts you in the certified talent pool, which micro1 is direct about: certification “does not guarantee immediate placement on a project.” From there you are matched two ways, by invitation as relevant roles open and by applying directly to postings in the portal. When a match lands, onboarding brings invitations to Slack, a micro1 expert email, access to the work platform, Hubstaff for time tracking, and Karma Check, the background check platform micro1 partners with, plus pipeline instructions, training materials and sample tasks to practice. micro1 publishes no retake policy for a failed interview and no typical wait between certification and a first paid task.
How it pays
micro1 publishes pay on nearly every role on its own job board: 313 of the 329 roles currently listed carry an hourly figure (4 list no pay, 3 list annual or monthly salaries). The median listing's midpoint is $70/hr, and the middle half of listings falls in the $40–$105/hr band. The full spread runs $6–$400/hr, but those ends are single outlying roles rather than anything typical. What we can't verify yet is how a listed rate translates into real, sustained take-home — that's what we're testing.
- $0 to $24 an hour: 45 listings.
- $25 to $49 an hour: 63 listings.
- $50 to $74 an hour: 64 listings.
- $75 to $99 an hour: 47 listings.
- $100 to $124 an hour: 47 listings.
- $125 to $149 an hour: 16 listings.
- $150 to $174 an hour: 11 listings.
- $175 to $199 an hour: 1 listing.
- $200 to $224 an hour: 1 listing.
- $225 to $249 an hour: 0 listings.
- $250 to $274 an hour: 18 listings.
micro1's experts guide publishes indicative bands by task type: $15 to $40 an hour for general annotation, $20 to $60 for prompt writing and RLHF evaluation, $50 to $100 for domain-specific review, and $100 to $200+ for frontier expert projects. It states that pay varies by project and verification level and is listed on each role when you apply, and two third-party trackers that read the board in September 2026 both report rates on individual listings, with AITraining.jobs logging advertised rates on 305 of 307 roles on September 12. Payment runs through Deel at payroll.micro1.ai on a semi-monthly cycle: work from the 1st to the 15th is paid around the 20th, work from the 16th to month end around the 5th, moving to the next business day when that date falls on a weekend or US bank holiday. Hourly pipelines are tracked in Hubstaff and payroll is calculated from exported Hubstaff data at cycle end, while pay-per-task pipelines are calculated from completions instead. US contractors can download Form 1099-NEC from their Deel account under Finance then Tax Forms, and micro1 warns that missing documentation or a wrong country of jurisdiction will delay payment.
What workers say
The public record on micro1
Trustpilot carries 10,184 reviews at a 4.7 TrustScore, and that pool is overwhelmingly people who had just finished the Zara interview, many of them before any outcome was known. Within it the sentiment is lopsided in micro1's favor, 83% five-star against 3% one-star, and the praise is consistent: the AI interviewer adapts, sounds natural, and is comfortable for non-native English speakers. The 1- and 2-star reviews from August 11 to September 11, 2026 cluster just as tightly, on interviews ending abruptly, disconnections and audio faults, too little time to answer, repetitive questions, unresponsive support, and a recurring suspicion that the interview exists to collect expertise rather than to hire. Almost nothing in the public review record is written by someone describing the paid work itself.
Scores, and what each one measures
- Trustpilot ↗4.7 out of 5
What it measures. micro1's AI screening interview, not the job. The pool is dominated by candidates reviewing the Zara interview right after taking it, many before they knew the outcome. Almost none of these reviewers describe being placed on a pipeline, doing paid tasks, or being paid. Read it as a rating of a recruiting funnel.
Our read of it
Pros
- Pay bands are published by task type before you go anywhere near an application: micro1's experts guide lists $15 to $40 an hour for general annotation, $20 to $60 for prompt writing and RLHF evaluation, $50 to $100 for domain-specific review, and $100 to $200+ for frontier expert projects.
- Rates appear on individual roles. micro1's guide says pay is listed on each role when you apply, and two independent trackers that read the board agree: AITraining.jobs logged advertised rates on 305 of 307 open roles on September 12, 2026, and Second Talent reported 326 roles with rates attached on September 11, 2026.
- The payment machinery is documented to the day. Deel, a semi-monthly cycle, the 1st to 15th paid around the 20th, the 16th to month end around the 5th, approvals around the 3rd and 18th, processed in PST, with weekend and bank holiday dates rolling forward. You can predict when money lands, which is rare in this category.
- There is a published route up. micro1 documents an advancement path from Expert to Reviewer to Human Data Manager to Human Data Lead to Strategic Project Lead.
- You can rehearse the screening for free. micro1's interview-prep page offers an 18-minute practice interview tailored to a selected role, with feedback, before you sit the real one.
- Referral payouts are flat amounts shown on each posting, up to $3,000, paid once your referral signs up through your link, certifies, is hired, and completes 10 hours of work or the task equivalent. You know the number before you refer anyone.
Cons
- Passing buys eligibility, not work. micro1 states that certification “does not guarantee immediate placement on a project,” publishes no expected wait to a first paid task, and says pipelines continue only as long as the client has work to be completed.
- Onboarding includes a third-party background check. micro1's onboarding page lists Karma Check alongside the Slack and email setup and describes it as the background check platform it partners with, with nothing on what is checked, what disqualifies you, or who sees the result.
- Hourly work requires the Hubstaff desktop tracker, and micro1's page on it covers creating the account, installing the app and clicking the timer. There is no published policy on screenshots, activity monitoring or idle time, so you are agreeing to a tracker whose settings are not documented anywhere you can read first.
- The contract is heavier than the marketing. Everything you produce is assigned to the company as Contractor Work Product, confidentiality covers the existence of the agreement and the identity of the client, a one-year non-solicit survives termination, and either party can end the arrangement on five business days' written notice without cause.
- Quality scoring has no published floor. micro1 says tasks are scored on review results and a pipeline lead will reach out if your performance needs improvement, but publishes no threshold, no removal procedure and no pay consequence.
- micro1's own pages disagree on how often you get paid. The experts guide says payments are processed biweekly, which is 26 cycles a year; the payroll policy describes a semi-monthly cycle, which is 24. Nothing on the site reconciles them.
- micro1 nowhere states whether onboarding, training or calibration time is paid. Its onboarding page describes training materials and sample practice tasks and says nothing about compensating them.
Who it fits, and who should skip it
Best for
- Credentialed specialists aiming at $100+/hr frontier work
- People who want the rate published before they commit
- Anyone comfortable being screened by a live AI interviewer, out loud
- Part-timers who can hold 15 to 20 hours a week when a pipeline runs
Who should skip it
Skip micro1 if you need income starting on a known date. Nothing it publishes promises work after you certify, the contract ends on five business days' notice from either side, and a pipeline lasts only as long as the client keeps buying. Skip it too if your background does not reach the upper bands, because the $15 floor for general annotation is not worth a 20 to 40 minute interview plus an onboarding sequence micro1 nowhere says it pays for, and skip it if being assessed by an AI on live spoken answers is a format you already know works against you.
Provenance
Where all of this came from
Confirmed on micro1's own site
Statements read directly on micro1's own pages.
- micro1's own job board publishes pay on nearly every role: 313 of the 329 roles currently listed carry an hourly figure (4 list no pay, 3 list annual or monthly salaries), with a median listing midpoint of $70/hr and the middle half in the $40–$105/hr band. Its largest specialist section is Law (41 priced roles).
- Every candidate is interviewed by Zara, micro1's AI recruiter, in a live session of open-ended spoken questions running 20 to 40 minutes, at roughly seven minutes per skill assessed.
- Payment runs through Deel on a semi-monthly cycle: the 1st-to-15th period is paid around the 20th and the 16th-to-month-end period around the 5th, with weekend and US bank holiday dates rolling forward.
- Either party can end the expert agreement without cause on five business days' written notice, all Contractor Work Product belongs to the company, and a one-year non-solicit survives termination.
Every source behind this review
What each one says, and how much weight it deserves.
Self-description as a “Data lab to train frontier models & evaluate agents,” with client product lines Realm (RL environments), Cortex (contextual evaluation) and robotics data.
How to read it. Company marketing about its own products. Treated as self-description throughout.
Four-step flow (apply, AI interview, onboard, start earning); fields listed as engineering, finance, healthcare and medical, legal, economics, linguistics, cinematography and more; early-career through PhDs; certification “does not guarantee immediate placement on a project.”
How to read it. Publishes no pay figures and no payment terms anywhere on the page.
The portal is a headline and a submission form rather than a browsable public list of roles with rates.
How to read it. Listings are not visible to an unauthenticated visitor, so per-listing rates are reported here from third-party trackers rather than from the board itself.
Indicative pay bands by task type ($15-$40, $20-$60, $50-$100, $100-$200+/hr); pay varies by project and verification level and is listed on each role when you apply; payments processed biweekly in USD.
How to read it. micro1's own indicative bands, not observed rates. Its “biweekly” wording conflicts with the semi-monthly cycle in its payroll policy.
Zara is micro1's AI recruiter and interviews every candidate; open-ended spoken questions; interviews run 20 to 40 minutes, about seven minutes per skill; applications go through the Opportunities Portal, where recruiters set skill requirements per role.
How to read it. Company documentation. Silent on retakes and on proctoring.
Certification means you are in the certified talent pool and eligible for pipeline opportunities, with access to the community and a talent dashboard.
How to read it. Eligibility, not placement. No retake policy and no proctoring rules published.
Certified experts are matched by invitation as roles open and by applying directly in the Opportunities Portal.
How to read it. No expected waiting time to a first paid task is stated.
Three work types: frontier evals, real-world robotics recording, and contextual evals built around a client's workflows.
How to read it. No concrete example of a single unit of work is published.
Onboarding brings Slack, a Google Workspace expert email, work platform access, Hubstaff, and Karma Check, described as the background check platform micro1 partners with, plus training materials and sample tasks.
How to read it. Says nothing about what the background check covers, and nothing about whether onboarding or training time is paid.
Tasks pass through multiple checks by the creator, peer reviewers, the operations team and sometimes the client, producing task quality scores; a pipeline lead reaches out if performance needs improvement.
How to read it. No published score threshold, no removal procedure, no stated pay consequence.
The Hubstaff desktop app tracks working time; you accept an invite, install it, pick your pipeline and start the timer.
How to read it. Entirely silent on screenshots, activity levels and idle-time thresholds.
Contracts and payment run through Deel at payroll.micro1.ai; Hubstaff hours do not sync automatically and are exported at cycle end; pay-per-task pipelines pay on completions; documentation and country of jurisdiction must be correct or payment is delayed.
How to read it. Does not mention tax forms at all.
Semi-monthly cycle: 1st to 15th paid around the 20th, 16th to month end around the 5th, approvals around the 3rd and 18th, processed in PST, US contractors +1 business day, weekend and US bank holiday dates roll to the next business day.
How to read it. Conflicts with the “biweekly” wording in micro1's experts guide.
Expected commitment is roughly 15 to 20 hours a week depending on client requirements, with more available on client demand; project duration varies by client demand.
How to read it. An expectation set by the company, not a guarantee of available hours.
Pipelines continue as long as the client has work; when one ends micro1 says it tries to map experts onto other pipelines; declining an offer has no bearing on future offers; simultaneous pipelines need both leads' approval; Form 1099-NEC is downloadable in the Deel account under Finance then Tax Forms.
How to read it. The reassurances about re-matching and about declining offers are company statements about its own behavior.
Published advancement path from Expert to Reviewer to Human Data Manager to Human Data Lead to Strategic Project Lead.
How to read it. A published ladder, with no data on how many people move up it.
Independent contractor status; either party may terminate without cause on five business days' written notice; all Contractor Work Product belongs to the company; confidentiality covers the existence of the agreement and the identity of the client; one-year non-solicit after termination; you warrant you are not currently employed by a foundational AI model company.
How to read it. micro1's own summary of the agreement rather than the signed document.
Up to $3,000 per successful referral, a flat amount shown on each posting, paid after the referral signs up through your link, certifies, is hired and completes 10 hours of work or the task equivalent.
How to read it. The $3,000 is a ceiling. Actual payouts vary by role.
Free 18-minute practice interviews tailored to a selected role, with feedback.
How to read it. Does not document recording use, proctoring or retakes.
TrustScore 4.7 across 10,184 reviews, 83% five-star and 3% one-star; recent reviews are dominated by candidates assessing the Zara interview.
How to read it. This rates the screening interview, not the job. Reviewers are applicants, most of them posting immediately after their interview.
Complaints from August 11 to September 11, 2026 cluster on abrupt endings, disconnections and audio faults, insufficient answer time, repetitive questions, unresponsive support, and suspicion that interviews harvest expertise.
How to read it. Self-selected negative reviewers, and these are allegations about the interview rather than about paid work.
307 open micro1 roles on September 12, 2026, of which 305 carry an advertised hourly rate.
How to read it. An automated scraper of micro1's feed. Used here only for the fact that rates appear on individual listings, not for its rate figures.
Reports 326 open roles on micro1's board on September 11, 2026, with hourly rates attached to individual listings.
How to read it. A competing commercial review site. Cited only as a second observation that listings carry rates.
Describes Zara as a GPT-4o-based system running practice interviews, conversational assessment, automated feedback and retrieval-based candidate Q&A.
How to read it. Written by micro1 about its own product. A company claim in an academic format, not independent evaluation.
Evidence scorecard
What we can back up about micro1
28
0
Not one of them is work we've done ourselves. Read every line below as what somebody else states, discounted by its tier.
- Tier 1 · Verified first-hand0 accounts
We did it ourselves and can describe it, with a date. No dated first-hand account of micro1's own process is on record here.
- Tier 2 · Platform-published23 statements
Stated on the platform's own site, board, or in writing to us. Read off micro1's own pages, and cited to them.
- Tier 3 · Worker reports0 reports
Said by people who say they've done the work. We hold none for micro1.
- Tier 4 · Unverified claim5 claims
Printed by a marketing or aggregator page with no sourcing behind it. We repeat these labeled so you can recognize them elsewhere, never as fact.
Listed pay
$70/hr median listed rateTier 2 · Platform-publishedThe median listing's own midpoint across the 313 of 329 open roles that publish an hourly figure, with the middle half between $40 and $105/hr. This is what micro1 advertises for a role, not what anyone has been shown to earn for working it.
Effective pay
Not knownNobody has published itWhat a person clears per hour of their life after unpaid application time, unpaid training, idle time between tasks and self-employment tax. Nobody publishes it for micro1, and we can't tell you yet either.
Getting in
Interview / vetted (per platform)Tier 2 · Platform-publishedmicro1's own material describes the step, and the review names the page it came from. What share of applicants clear it, micro1 does not publish.
Specialty labels
micro1's own sectionsTier 2 · Platform-publishedEvery listing carries the section micro1 filed it under. We count them; we don't decide them.