Mise à jour Octobre 17, 2026
Mercor AI connects professionals with AI-related contract work and provides expert-led evaluation services for companies. Its current offering extends beyond recruiting: the site web officiel also highlights human data, APEX benchmarks and enterprise AI deployment. This overview explains the public offering and application requirements; it is not a hands-on test or a promise of employment.
What does Mercor offer?
For professionals, Mercor describes remote projects that can involve reviewing model responses, creating specialist examples and assessing output quality. Relevant fields include software engineering, science, finance, law and medicine. Requirements vary by role; some call for substantial experience or advanced qualifications. See the expert overview for the current positioning.
Pour les entreprises, la enterprise evaluation service distinguishes three approaches:
- Expert staffing: Mercor sources and vets specialists who work under the client’s direction.
- Production rubrics: expert-designed criteria assess agent behavior against real operational traffic.
- APEX benchmarks: structured, offline environments test performance on professional tasks.
APEX is now a family of evaluations rather than one interchangeable score. For example, the homepage distinguishes single-turn APEX-1 tasks from cross-application APEX-Agents work. Compare the task domain, tools and evaluation method before using any benchmark result to judge suitability for your workflow. Mercor also advertises enterprise AI deployment with embedded engineering support; a benchmark score alone does not establish production reliability.
How the expert application works
Le guide d'application describes choosing an opportunity, uploading a résumé, completing the required assessments and confirming work authorization. Some completed assessments can carry over to compatible applications. The exact steps depend on the role, and completing them does not guarantee an offer.
AI interviews ask role-specific questions and generate a transcript and evaluation. Mercor’s guide d'entretien recommends current Chrome on a computer; Edge and Safari are also supported, while mobile browsers and Firefox are not supported for interviews. Camera and microphone access are needed, and some assessments require screen sharing. Browsing opportunities on a phone is different from completing an interview.
A hypothetical preparation workflow: a financial analyst could check a listing’s qualifications and location restrictions, update their résumé with accurate experience, test their camera and microphone, and use a practice interview to become familiar with the interface. Before screen sharing, close unrelated documents and private notifications. Answer independently and follow the assessment’s AI-use and confidentiality rules. This is preparation guidance, not a technique for securing a particular score or offer.
Pay, contract types and company pricing
Expert compensation is project-specific. Mercor says rates depend on expertise, project complexity and demand. Check the actual offer rather than relying on headline earnings or a platform-wide average. Projects can be extended, shortened or ended early, so neither continuous work nor a particular income should be assumed.
Le contract guide distinguishes hourly work from deliverable-based contracts. Hourly contracts use time tracking through Insightful and may have weekly caps. Deliverable-based work pays for successful completion of specified output and does not guarantee a minimum hourly return; most such contracts do not require time tracking unless stated otherwise. Complete onboarding before beginning work, and check acceptance requirements, deadlines and payment terms.
Mercor documents weekly processing for hourly contracts, but receiving funds also depends on submitted hours, payment processing and account setup. Its prévisions de revenus explains payment status and payout arrangements. This should not be read as a guaranteed arrival date.
For companies, no universal public price was established in the pages reviewed. Mercor’s staffing documentation describes a margin on company pay rates; its enterprise page also offers task-based pricing for end-to-end evaluations. Request a scoped quote covering specialist requirements, deliverables, acceptance criteria, data rights and total cost rather than assuming every engagement uses the same hourly fee.
Privacy and practical limitations
Le AI and data-use guide describes collecting résumés, recorded interview audio/video and transcripts, and using profile information for matching. It also describes internal recommendation-model training while saying interview data is not used to train external AI models. Another interview FAQ uses broader no-training wording, so a blanket “your data never trains AI” claim would be misleading. Turning off job outreach does not hide your profile from companies.
Le Avis de confidentialité du site Web, updated August 26, 2026, explicitly excludes the service at work.mercor.com and points to an Expert Privacy Notice. Read the applicable expert notice during signup or ask support for it before submitting information. Account-deletion guidance also lists restrictions for active contracts, pending payouts and unanswered offers; legal retention and copies already accessed by companies may remain.
- AI screening has limits: an automated assessment is not proof of objective, bias-free selection or a guarantee of hiring.
- Contract details matter: review confidentiality, ownership of work, eligibility and tax obligations before accepting. Get qualified advice where needed.
- Use professional judgment: evaluating legal, medical or financial AI output is distinct from authorizing that output for real-world decisions.
- Device requirements vary: interviews need a supported computer browser, and hourly work can require a desktop tracking application. Do not assume a phone-only workflow will work.
Who is Mercor worth considering for?
Mercor may suit specialists willing to complete assessments and work on variable contract projects, and organizations seeking expert input into AI evaluation. It is a weaker fit for someone needing guaranteed hours, predictable long-term earnings or fully mobile participation. Start with a specific role or a small, clearly defined client evaluation project, and verify its terms before committing. For other categories of software, browse our Répertoire des outils d'IA.
