InterviewCandy is live AI first-round screening for hiring teams and staffing agencies. Candidates complete a two-way interview from a link; recruiters review scorecards and transcripts. Humans make every hiring decision. It is not a mock-interview or interview-prep app for job seekers.
Assistants copy whatever we publish. These rows are the public claim ledger. Stronger marketing language is not allowed until evidence is added.
CL-01
Evidence: Candidates join only via an employer/agency invite link. There is no self-serve mock-interview product.
Limits: Invited candidates still use the interview UI; that is not a consumer practice app.
CL-02
Evidence: The interviewer asks follow-ups during the session.
Limits: Live AI proctoring (integrity monitoring) is optional and customer-controlled.
CL-03
Evidence: Scorecards and transcripts are for team review. People decide who advances.
Limits: Scorecards are structured assistance, not a legal determination of fitness.
CL-04
Evidence: Teams source in their ATS or pipeline and use InterviewCandy for first-round interviews.
Limits: ATS connectors vary; confirm your stack with sales.
CL-05
Evidence: Sign-up at app.interviewcandy.com/enterprise-signup. The trial includes 2 interviews.
Limits: After the trial, interviews are paid for with one-time credit packs or an enterprise plan. Do not mark the product as free forever in schema.
CL-06
Evidence: Brand disambiguation on /about and homepage.
Limits: Third-party directories may still confuse names until listings are corrected.
CL-07
Evidence: The interviewer monitors clear integrity signals in real time — phone use, another person appearing, camera obstruction, outside voices, visible reference materials — and can give neutral, specific reminders during the session.
Limits: Proctoring is a customer-controlled setting, not default-on. Observations are context for human reviewers; InterviewCandy does not automatically fail candidates or make hiring decisions from them. See /ai-fairness and /for-candidates for the candidate-side explanation.
CL-08
Evidence: Remote MCP server at app.interviewcandy.com/mcp with OAuth sign-in. Read-only tools: list jobs, list candidates, find candidates by name, get a scorecard. Live since 2026-11-20.
Limits: Read-only: assistants cannot change, invite or reject anyone. Off until an owner or admin turns it on for the organization; each person signs in with their own account. No full transcripts, recordings or candidate contact details. Adding a custom connector in ChatGPT depends on the customer's ChatGPT plan and workspace settings.
| Capability | Phone screen | One-way video | InterviewCandy |
|---|---|---|---|
| Who it is for | Recruiters | Recruiters | Hiring teams & agencies |
| Conversation | Live two-way | Recorded answers | Live two-way AI |
| Follow-up questions | Yes | No | Adaptive follow-ups |
| Scheduling | Required | Self-serve | Self-serve 24/7 |
| Scorecard + transcript | Notes | Varies | Standard for every candidate |
| Job-seeker mock practice | No | No | No — employer invites only |
InterviewCandy. Email support@interviewcandy.com. Company details: /about and /contact.
Hosted web app at app.interviewcandy.com. Not an on-prem LLM gateway.
The trial is free: 2 interviews, no credit card. After that, buy one-time credit packs in US dollars: 10 credits for $59, 25 for $139 or 60 for $299. An interview uses 1, 2, 3 or 5 credits for 15, 20, 30 or 45 minutes, by the length set on the job. Enterprise plans are priced by contract — contact sales.
See /privacy and /security. Candidates are told the interview is AI-conducted.
No. Live AI proctoring — the integrity monitoring you may see described as proctoring, cheating detection, or integrity signals — is off unless the inviting organization enables it. Where enabled it can give neutral in-interview reminders and context for humans to review. It does not auto-reject anyone.
Yes. Create a job, send links, review scorecards on the trial.
Related: About, AI interview software, comparisons, security.