Skip to content

Platform · The MirageScouting Engine

Every candidate measured on the same evidence.

The MirageScouting Engine runs under Scout, Academy and Office. It interviews by voice, scores each skill from Beginner to Master level with the candidate's own words as evidence, and leaves the decision to your team.

  • Levels calculated from evidence
  • English & German, up to 70 languages
  • Three interview rounds
  • Recommend-only
  • REST API & webhooks (early access)

One engine · three products

  • MirageScoutingHiring
  • MirageAcademyLearning
  • MirageOfficePeople Analytics

MirageScouting Engine

  1. Skill frameworks
  2. MirageScouting Voice Engine
  3. Evidence-based evaluation
  4. Verified Capabilities
Scout, Academy and Office build on the same skill frameworks and the same profile. MirageScouting Voice Engine runs the voice interviews in Scout and the coaching in Academy.

The engine

What sits under every MirageScouting product.

Recruiters see interviews and reports. HR leaders see data they can compare across roles and teams. Both come from the same engine: every product shares the skill frameworks, the evaluation and the profile, and MirageScouting Voice Engine runs the voice interviews and the coaching.

Skill frameworks: the yardstick

Each role is broken down into skills with level descriptions from Beginner to Master. Our product management framework covers 40 skills in 6 categories; frameworks for further roles are set up with our team.

MirageScouting Voice Engine: the conversation

The voice engine holds spoken interviews in the language you set per job, starting with English and German, and follows up when an answer stays general. Academy coaching sessions run on it too.

Evaluation: the evidence

Each level is calculated from the candidate's own words in the interview. Claims from the CV and the interview are marked as confirmed, unverified or contradicted.

One profile: the record

Assessments and reviewed interview results build one skill profile per person, so the next round and the next role start from what is already known.

One engine, three products

  • MirageScoutingHiringJob posts, structured AI interviews, the pipeline board, talent scouting and twelve hiring metrics.Scout for companies
  • MirageAcademyLearningLessons, voice coaching on MirageScouting Voice Engine and certification programmes, measured on the same skill frameworks.Explore the Academy
  • MirageOfficePeople AnalyticsComing soonEmployee profiles, feedback, pulse surveys and performance reviews, on the same skill profile as Scout.

Decision engine & MCP

Decisions on evidence. Discoverable in ChatGPT, Claude and Gemini.

Every interview follows a protocol, every level is calculated from the candidate's own words, and a person on your team makes every decision. Our MCP service connects MirageScouting to ChatGPT, Claude and Google Gemini, so candidate and company profiles get discovered where people and AI agents already ask.

How a decision is made

  • Decision engineA model only finds the candidate's exact words. The level is then calculated in code: evidence for a level raises the floor, contradicting evidence caps it, and what was never asked only widens the range. When the evidence does not settle one level, you see a range, not a guess. The engine recommends; a person on your team decides.
  • Interview protocolsThree rounds with fixed time boxes: pre-screening in about 30 minutes, technical and collaboration in up to 75. The questions come from a reviewed bank of more than 850 in English and German, aimed at the higher of the claimed and the required level, each with a follow-up for answers that stay general. Every question the engine picks is recorded.
  • StandardsOne scale from Beginner to Master for candidates, companies and the Academy, with more than 2,600 observable level indicators. Consent comes before every interview, there is no emotion or tone analysis, and a person approves a level before it reaches a profile.

We do not train AI models on candidate or customer data.

MCP: MirageScouting inside ChatGPT, Claude and Gemini

Live

Company profiles and open roles

Ask ChatGPT, Claude or Gemini about employers and open roles. They find company profiles and live roles on MirageScouting and answer with the details.

Live

Candidate profiles by skill level

Recruiters and AI agents search the profiles of candidates who opted in, by skill and minimum level, from Beginner to Master, inside the assistant they already use. Candidates get found for what they can do.

Consent and control

Opt-in only: companies and candidates decide whether they appear, and candidates start switched off. Single roles can be hidden, contact details stay private and every agent request is logged.

Interview automation

Three interview rounds. You decide how much the AI does.

Switch on the rounds a role needs. Every candidate in the role gets the same structure, the same rubric and the same report format.

The rounds

  1. Round 1~30 minPre-screeningMotivation, current situation, ways of working, salary expectations and availability. You pick the topics.
  2. Round 2up to 75 minTechnical & hiring managerThe role's skills in depth, with case work such as a technical architecture case or a 30/60/90-day plan.
  3. Round 3up to 75 minCollaborationHow the candidate works with others: the soft skills, behaviours and business understanding you choose to observe.

Two ways to run a round

Live

Full AI

MirageScouting Voice Engine runs the round end to end: invitation, consent, interview and report. Candidates choose the moment; your team reads the evidence and decides.

Coming soon

AI Assisted

Your recruiter leads the conversation in your own video call. MirageScouting Voice Engine prepares the question plan, suggests the next question during the call and writes the evidence-backed report afterwards.

AI Assisted Interviews are coming soon. Then you set the mode per round: Full AI for pre-screening volume, a person in the room for the hiring-manager round. Either way, MirageScouting never moves an application on its own.

Talk to us about early access to AI Assisted Interviews

MirageScouting Voice Engine

MirageScouting Voice Engine: the interviewer that follows up.

MirageScouting Voice Engine is the voice layer of the MirageScouting Engine. It turns your role setup into a spoken interview that probes like an experienced interviewer and stays inside the plan your team approved.

MirageScouting Voice EngineVoice Engine V4.26

A plan for every candidate

Before each interview, MirageScouting Voice Engine builds the question plan from the round, the role, the seniority, the candidate's skill gaps and the evidence from earlier rounds.

Follow-up where it matters

When an answer stays general, MirageScouting Voice Engine asks for a concrete example, so the score rests on something specific.

Inside the time box

Around 30 minutes for pre-screening, up to 75 for deeper rounds. Lower-priority questions give way before an interview overruns.

Up to 70 languages

Interviews run in the language you set per job. We start with English and German and add languages fast, up to 70. For roles in Germany, MirageScouting Voice Engine also asks about notice period and salary expectations.

Your team configures, per job

  • Rounds
  • Skills & minimum levels
  • Required skills
  • Interview language
  • Interviewer persona
  • Pre-screening topics
  • Case types
  • Soft skills
  • Behaviours
  • Business understanding

MirageScouting Voice Engine does not score voice, accent, appearance or emotion. Scores come from what the candidate says, measured against the skills you set.

API & ATS integration

Run MirageScouting headless, from the systems you already use.

Start interviews, follow their status and fetch reports through our REST API. Your ATS stays the system of record; MirageScouting adds the interview and the evidence.

Early access

Your IDs, not ours

Every call uses your own job, candidate and interview IDs. You never have to store a MirageScouting ID.

Scoped, revocable keys

Each key carries only the scopes you grant: records, interviews, reports or employee sync. The organisation comes from the key, never from the request.

Signed webhooks

MirageScouting tells your endpoint when an interview is completed, its report is ready or it is cancelled. HMAC-signed, free of assessment data, retried for almost a day.

No candidate account

Candidates get a personal, revocable link, accept the consent notice and interview without signing up.

The interface

  • POST/ats/v1/jobsSend a job from your ATS
  • POST/ats/v1/candidatesRegister a candidate
  • POST/ats/v1/interviewsRequest an interview: round and mode
  • GET/ats/v1/interviews/{externalInterviewId}Follow its status
  • GET/ats/v1/interviews/{externalInterviewId}/reportFetch the report
  • DELETE/ats/v1/interviews/{externalInterviewId}Cancel and revoke the link
  • POST/ats/v1/employeesSync your employee roster to Office

Webhook events

  • interview.completed
  • interview.report_ready
  • interview.cancelled

Works with any ATS that can call a REST API and receive webhooks. There are no pre-built connectors for specific vendors yet.

Ask about early access

Across the platform

Reach, security and market data, on the same foundation.

Media & reach

Put your roles in front of the right people.

Media & reach

Promotion options increase how many people see your roles and your company. They never change where a candidate ranks.

  • Sponsored boosts and featured roles, labelled as sponsored
  • Newsletter features and a featured company profile

Security

Your data, protected in every layer.

Security details

Databases in Germany, TLS for connections over the internet, and strict separation between organisations.

  • Databases in Germany (EU)
  • Connections over the internet use TLS. Files such as interview transcripts are encrypted at rest by our storage provider.
  • Each organisation sees only its own jobs and candidates
  • SOC 2 planned

Market analytics

Plan roles against the market.

Market analytics

Salary bands, skill demand and remote share for a role and level, from the roles posted on MirageScouting and market research, next to the metrics of your own pipeline.

  • Salary bands by seniority: 25th percentile, median, 75th percentile
  • The most requested skills and how they relate to pay
  • Twelve hiring metrics, from time to hire to source effectiveness

FAQ

What HR leaders ask about the engine.

What is the MirageScouting Engine?

The shared system under MirageScouting, Academy and Office: role skill frameworks with levels from Beginner to Master, evidence-based evaluation and one skill profile per person, plus the MirageScouting Voice Engine for interviews (Scout) and coaching (Academy). Because every product uses the same frameworks and scale, results stay comparable.

Which AI models does MirageScouting use?

Google Gemini for the live conversation, Anthropic Claude for question plans and evaluations, and OpenAI for indexing role documents and as an alternative real-time voice. We do not train AI models on candidate or customer data.

Does the AI decide who gets hired?

No. MirageScouting recommends advance, hold or reject, with the evidence. Only a person on your team moves an application. That is how the product works, not a setting.

What is MirageScouting Voice Engine?

MirageScouting Voice Engine is our voice engine, currently V4.26. It holds the spoken interview, follows up on vague answers, keeps to the time box and also runs Academy coaching sessions.

Which interview languages are available?

Interviews run in English and German today, and we are adding languages fast, up to 70. You set the language per job, and for roles in Germany MirageScouting Voice Engine also asks about notice period and salary expectations.

How much of the interview can we customise?

Per job you choose the rounds, the skills and minimum levels, the interview language, the interviewer persona, the pre-screening topics, the case types and the soft skills to observe. Frameworks for roles beyond product management are set up with our team.

Can we keep our ATS?

Yes. API access is in early access: your ATS requests interviews, receives signed webhooks and fetches the reports. There are no pre-built vendor connectors yet.

Which AI assistants work with MirageScouting over MCP?

ChatGPT, Claude and Google Gemini. Assistants and AI agents discover company profiles and open roles, and the profiles of candidates who opted in, by skill and level. Nothing is shared until a company or candidate switches it on.

How do we start with AI Assisted Interviews?

AI Assisted Interviews, where your team leads the conversation, are coming soon. Book a discovery call through the contact form (topic: AI Assisted Interviews) to hear when they open and to get early access.

See the engine on one of your open roles.

Thirty minutes with your job description: we set up the rounds, walk through a report and show where the API fits your stack.