RoundHound
HomeSign in

Privacy Policy

This policy explains what data RoundHound handles when you research a role, practice, use voice or mentor rooms, and buy a pass.

Last updated November 3, 2026

On this page

  1. 01Scope and operator
  2. 02Data we collect
  3. 03Shared research boundary
  4. 04How we use data
  5. 05AI and voice processing
  6. 06When data is shared
  7. 07Payments
  8. 08Cookies and analytics
  9. 09Retention and deletion
  10. 10International processing
  11. 11Security
  12. 12Your privacy rights
  13. 13Children
  14. 14Updates and contact

Questions about this document?

[email protected]

Scope and operator

This Privacy Policy explains how RoundHound, the operator of the Service, collects, uses, discloses, and retains personal data when you visit, create an account, research a target, practice, use voice features, invite a mentor, apply as an interviewer, buy a pass, or contact us. It should be read with our Terms of Service.

The current privacy contact is [email protected]. If another organization gives you access to RoundHound or submits data about you, that organization may have its own privacy responsibilities as well.

We do not sell personal data for money. When we advertise, we share limited website activity with advertising platforms to measure our own ads. Some U.S. state laws treat this as “sharing” for cross-context behavioral advertising. You can opt out with Global Privacy Control or the control under Cookies and analytics.

Data we collect

Account and identity

We process your email address, display name, sign-in provider and provider identifier, email verification state, account role and status, profile preferences, locale or time zone where supplied. Google/Firebase handles password and Google sign-in credentials, and Apple handles Sign in with Apple in the iOS app; if you choose to hide your email there, we receive Apple's private relay address instead. RoundHound receives identity tokens and keeps an application account and secure session records.

Targets, profile, and files

You may provide a company, role, seniority, location, team, interview date, job-posting URL or text, recruiter notes, preparation preferences, personal context, résumé, story-bank content, or questions. An uploaded résumé PDF is parsed during upload. The app keeps its filename and extracted text; it does not store the original PDF. We also store the plans, research references, rounds, rubrics, and generated materials associated with your account.

Rehearsals and feedback

We process session status and timing, typed and transcribed spoken messages, AI interviewer turns, private coaching conversations, code, diagrams, whiteboard content, documents and notes, workspace changes, question feedback, reviewer notes, assessments, loop reports, progress signals, and the actions needed to save and resume a room. Coaching can be kept out of the assessed interview evidence while still being stored as part of your session.

Mentor and interviewer activity

For mentor rooms, we process invite codes, candidate and mentor account identifiers, display names, request and attendance state, the shared room record, mentor feedback, and any mentor-assistant state used in that room. If you apply to be an interviewer, we process the profile, experience, availability, links, motivation, review decision, and activity needed for that role.

Technical, security, and support data

We process request and correlation identifiers, session and CSRF tokens in protected form, IP and user-agent information or hashes, sign-in and account activity, audit events, rate-limit and abuse signals, error details, provider/model identifiers, token or speech usage, latency, job state, and internal cost records. If you contact us, we process the message, contact details, attachments, and follow-up needed to respond.

The shared research boundary

Public-company research is deliberately separated from personal tailoring. RoundHound may reuse a research dossier across users when they ask about the same public target. That dossier can be scoped by normalized company, role, seniority, location, team, a sanitized public job-posting URL, and an opaque hash of supplied posting text. It can contain public sources, public job facts, research summaries, likely round patterns, and evidence notes.

Your résumé, personal notes, recruiter notes, full pasted job text, transcript, workspace, and private practice preferences are excluded from the shared research dossier. They are used later to tailor your own plan or session. URL credentials, fragments, and unrelated tracking parameters are removed before a job-posting URL enters shared research; only limited job-identity parameters may remain.

Because this cache is based on public target information rather than account ownership, it may be retained, refreshed, and reused independently of the account that first requested it. If a public source or posting contains personal information that should not appear there, contact us so we can review it.

How we use data

We use personal data to:

  • create and secure accounts, verify identity, maintain sessions, and enforce account roles;
  • research targets, interpret job information, parse résumés, and create tailored plans and original practice material;
  • run text or voice rehearsals, save transcripts and workspace state, provide coaching, and generate feedback and progress views;
  • create human interview rooms, honor the open invite link you share, and operate interviewer workflows;
  • provide the free round and free trials, create and verify subscriptions and one-time pass orders, prevent duplicate grants, and administer refunds or disputes;
  • respond to support and privacy requests and communicate important service, account, security, or payment information;
  • detect abuse, account sharing, fraud, attacks, unsafe content, and violations of our Terms;
  • debug, monitor, measure, and improve reliability, usability, research quality, and AI cost controls; and
  • comply with law, enforce agreements, preserve legal claims, and protect users, providers, RoundHound, and the public.

Where a law requires a legal basis, these activities rely as applicable on performing our agreement with you, our legitimate interests in operating and protecting the Service, your consent where requested, and compliance with legal obligations. You can contact us about the basis for a particular use.

AI and voice processing

AI features

Research, intake, plan generation, question creation, AI interviewing, coaching, workspace interpretation, transcription, and review can send relevant inputs to AI or search providers. Depending on the task, this may include public target details, a limited research brief, your résumé or notes, session transcript, prompt, workspace evidence, and generated material. We separate web-enabled public research from private résumé and session personalization so private content is not attached to the public web-research step.

In the iOS app, we ask for your permission before any of this is sent to these providers. If you don't agree, the app signs you out and sends nothing.

The provider used can depend on the task and configuration. Provider handling is governed by our configuration, our arrangements with them, and their applicable terms and policies; this Policy does not make additional data-use promises on a provider's behalf.

Free tools

The free tools under /tools and the answer boxes in our guides need no account. What you say or type there is sent to the same AI and speech-recognition providers to give you feedback, and we don't save it. We keep usage records (which tool ran, how long an answer was, and what it cost), and we briefly use your IP address to limit how often the tools can be used.

Voice features

When you turn on voice, microphone audio is streamed to the Service and a speech-recognition provider so it can produce live or completed text. Interviewer text may be sent to a speech- synthesis provider to return audio. Today these are Deepgram or Cartesia for speech recognition and Cartesia for speech synthesis. RoundHound does not retain the raw microphone audio.Audio is relayed for processing and is not written to our persistent storage. The resulting transcript is retained with the session just like a typed answer, and speech usage and request state may be logged without the raw recording.

Spoken transcripts can contain mistakes. You decide whether to enable the microphone and can use text mode where available. Browser and operating-system permissions control microphone access.

When data is shared

We disclose data only for the purposes described here:

  • At your direction: a mentor or interviewer you invite can receive the identity, target, prompt, transcript, workspace, timing, and feedback made available in that room.
  • Gift links: if you share your gift link, anyone who opens it sees your first name and whether a gift week is still available. When someone claims a week, we record which account gave it and which account claimed it so we can apply the gift limits and prevent abuse. Neither of you can see the other’s practice.
  • Infrastructure and security: cloud hosting, content delivery, database, backup, logging, and security providers process data needed to deliver and protect the Service. The current architecture uses Cloudflare and Amazon Web Services.
  • Identity and product measurement: Google/Firebase supports authentication and, when enabled, product analytics. Apple supports Sign in with Apple in the iOS app.
  • Email: Amazon Simple Email Service delivers our emails to your address. Firebase creates the one-time links in verification and password-reset emails.
  • Advertising measurement: when we run ads, Google Ads, Reddit, and Meta receive limited information about visits, sign-ups, and purchases, as described under Cookies and analytics.
  • AI, web research, and speech: AI and search providers such as OpenAI and, when configured, OpenRouter and its routed model providers process task inputs. Voice providers, currently Deepgram and Cartesia, process speech recognition and synthesis.
  • Payments: Stripe processes Checkout, payment methods, refunds, disputes, fraud controls, and payment communications. In the iOS app, Apple processes App Store purchases and refunds.
  • Legal and safety: we may disclose information when reasonably necessary to comply with law or legal process, enforce our Terms, investigate fraud or security incidents, or protect rights and safety.
  • Business change: information may be reviewed or transferred as part of a financing, reorganization, sale, merger, or similar transaction, subject to appropriate confidentiality and applicable law.

Service providers and exact models can change as the Service develops. They receive the categories of data reasonably needed for their function. We may also use aggregated or de-identified information that is not reasonably linkable to you.

Payments

Paid access uses Stripe-hosted Checkout for the monthly plan and one-time passes, and Stripe’s customer portal for changing a card or cancelling. Stripe collects payment and billing details directly. RoundHound stores the Stripe customer, Checkout Session, subscription, invoice and Payment Intent references, offer and pass version, trial and renewal dates, amount, currency, environment, status, entitlement, and any refund or dispute records needed to verify and administer access. We do not receive or store your full payment-card number through this flow.

In the iOS app, passes are bought through the App Store. Apple collects payment details and handles refunds. RoundHound receives the signed App Store transaction for the purchase (its identifiers, product, price and currency, purchase date, and environment) to grant your pass, and Apple's notifications of refunds or revocations so it can withdraw a refunded pass. We never receive your Apple payment details.

We retain transaction and entitlement records as needed for access integrity, reconciliation, fraud prevention, disputes, accounting, tax, audit, and legal obligations even if other account content is deleted.

Cookies and analytics

RoundHound uses essential cookies and similar storage for sign-in, secure application sessions, request protection, redirect state, and service continuity. Blocking them may prevent account features from working.

In staging or production, we may enable Firebase Analytics to measure a limited product funnel and feature use. Events can include actions such as landing-page and guide visits, sign-in success, target creation, round completion, review viewing, and pass checkout. When signed in, analytics can use a pseudonymous account UUID rather than your email, along with device or browser information and analytics identifiers supplied by the provider. Local and test environments do not send these events to that analytics service.

Cloudflare Web Analytics may also be enabled at the hosting edge. Its browser beacon measures page-load performance and traffic dimensions such as host and path, referring site, country, device type, browser, operating system, and navigation type. This measurement is separate from the product events sent to Firebase Analytics.

A first-party cookie notes how you first reached RoundHound: the referring site, any campaign in the link, and the page you landed on. It stays on your browser for up to 90 days, and if you create an account we save those details with it to learn which channels help people. We do not share them with advertising platforms, and they are deleted with your account.

When we advertise on Google, Reddit, or Meta (Facebook and Instagram), we measure which ads bring people to RoundHound. Google Ads can receive sign-up and purchase events from our Google Analytics data. Reddit’s pixel can record that a browser visited a page, signed up, or bought a pass, along with cookie identifiers, the page address, and device and browser information. Meta’s pixel can record that a browser visited a page, and our servers tell Meta when that browser signs up or buys a pass, with Meta’s cookie identifiers, a scrambled account identifier, the purchase amount, and browser information. These platforms may use this information to measure and improve ad delivery, including showing RoundHound ads to people who visited the site. Google Analytics also uses Google signals, which reports aggregated demographics and recognises visits across devices for people signed in to their Google account. We do not send these platforms your email address, résumé, interview content, or transcripts.

To opt out on this browser, turn on Global Privacy Control or use the control below. Opting out stops the Reddit and Meta pixels and our sign-up and purchase reports to Meta, and tells Google not to use your analytics data for advertising, from the next page you load. Essential cookies and product analytics continue.

Our emails: we send account emails (verification and password reset) and, unless you unsubscribe, a welcome email, a reminder after a week without practice, a note when a round's feedback is ready, and a note after your first pass purchase. Welcome, reminder, feedback, and purchase emails record whether they were opened, using a small image, so we can tell which emails are useful. Account emails are never tracked. Every welcome, reminder, and feedback email has an unsubscribe link; account and purchase emails still arrive after you unsubscribe.

Retention and deletion

We keep account data and user content while needed to provide and secure the Service and for the purposes described in this Policy. Retention depends on the type of record, whether your account or pass remains active, your deletion choices, operational and security needs, unresolved support, payment or legal issues, and applicable recordkeeping duties. We do not promise one fixed period for every category.

  • Plans, sessions, transcripts, workspaces, reviews, and résumés can remain while your account is active. Available product controls let you delete supported plans, sessions, and résumés.
  • Raw voice audio is not retained by RoundHound; the transcript is kept under the session's retention and deletion rules.
  • Shared public research can remain and be refreshed independently of your account, as explained above.
  • Security, audit, anti-fraud, payment, tax, allowance, and entitlement records may be retained after content deletion where needed for integrity, dispute handling, or law.
  • Deleted data may remain in protected backups until those backups age out, unless a legal hold applies.

Product controls let you delete supported résumés, sessions, and plans. In the iOS app you can also delete your whole account: your jobs, sessions and transcripts, feedback, résumés, and profile are erased right away, and your email and sign-in are removed from the account record. Payment, pass, and allowance records stay, no longer linked to your email. You can also contact us for account deletion or a broader privacy request. Deleting a practice record does not restore already-used time or allowances. We may verify identity and retain or pseudonymize limited records where law or service integrity requires it.

International processing

RoundHound and its providers may process data in countries other than the one where you live. Cloud infrastructure, identity, AI, speech, analytics, support, and payment providers may operate across multiple regions, and their laws may differ from yours. The transfer mechanisms and safeguards available depend on the provider and applicable law. Contact us if you need more information about a transfer relevant to you.

Security

We use technical and organizational measures designed to protect data, including encrypted transport, secure and HTTP-only application sessions, access controls, account-scoped ownership checks, hashed session and selected network identifiers, private application storage, audit events, provider-secret controls, rate limits, and encrypted backups. Payment access is granted only after verified provider events, not from a browser redirect.

No internet service is completely secure. Protect your account, use a unique password where you sign in by password, keep mentor links private, and contact us if you suspect unauthorized use. Do not send passwords or full payment-card details by email.

Your privacy rights

Depending on where you live, you may have rights to access, correct, delete, or obtain a copy of personal data; restrict or object to certain processing; withdraw consent where processing relies on consent; and complain to a data-protection or consumer authority. You may also have a right not to receive discriminatory treatment for making a privacy request. You can opt out of sharing for advertising as described under Cookies and analytics.

Send a request to [email protected]. Describe the right you want to exercise and the account concerned. We may ask for information reasonably needed to verify identity and authority, especially where somebody acts for you. We will respond under the process and timing required by applicable law. Some rights have exceptions—for example, records may need to remain for payment disputes, security, legal claims, or compliance.

Children

The Service is not directed to children under 16, and we do not knowingly invite them to create accounts. If you believe a child under 16 has provided personal data to RoundHound, contact us so we can review and take appropriate action. A higher local minimum age or requirement for parental authorization continues to apply where relevant.

Updates and contact

We may update this Policy as the Service, providers, or legal requirements change. The latest version will be posted here with a revised date. If a change materially affects how we use personal data, we will provide reasonable notice through the Service or available account contact information where required.

The current privacy contact for RoundHound is [email protected]. Please do not include a password, full card number, or unnecessary sensitive content in your first message.

RoundHound

Prepare thoughtfully. Show your own work.

Terms of ServiceSign inHome