Questions deserve direct answers.

ResidencyLens stores the account information, profile evidence, CV-derived fields, preferences, application work, and workspace progress needed to provide its features. Signed-in workspace records are associated with the authenticated user so the same evidence can power Dashboard, Strategy, Application, Interviews, Ranking, and Ava.

No. ResidencyLens does not sell applicant profiles, CVs, or workspace content to advertisers, programs, or data brokers. Your information is used to operate and improve the product experience you request.

No. The deterministic recommendation engines use your saved evidence as input at runtime; they do not retrain themselves on your CV. AI-assisted features may receive a bounded subset of relevant context to answer your request, but that is separate from training the ranking engine.

When you explicitly use Ava, document drafting, or an agent-assisted audit, ResidencyLens may send the question and the minimum relevant profile, strategy, or application context to OpenAI. Requests use storage-disabled settings where supported. Deterministic scoring and ordinary navigation do not require an LLM call.

Do not enter patient names, MRNs, dates or locations that could identify a patient, confidential letter text, credentials you are not authorized to share, or unnecessary sensitive information. The privacy check is a safeguard, not a guarantee.

ResidencyLens uses Supabase authentication, verified email, server-managed sessions, and protected cookies. Passwords are handled by the authentication provider rather than stored in readable form by the application. No web service can promise zero risk, so use a unique password and protect your email account.

Self-service account export and deletion are not yet available in the current beta. This is a known launch limitation, not a hidden capability. Until those controls are released, avoid storing information you do not need for the planning workflow.

No. A strategy-fit score is a planning index used to order and explain options. It is not the probability of an interview or Match, and ResidencyLens does not have access to each program's full screening process or applicant pool.

The deterministic strategy engine compares documented applicant evidence with specialty-specific applicant tiers, program rank tiers, fit bands, eligibility evidence, geography, preferences, and uncertainty. The apply list emphasizes a realistic portfolio; the signal shortlist uses a separate decision rule for limited expressions of interest.

They are decision support, not ground truth. Accuracy depends on the completeness of your profile and the quality, freshness, and coverage of program data. ResidencyLens exposes confidence and missing evidence so uncertainty is visible instead of being hidden inside a precise-looking number.

Yes. Program cards separate fit, portfolio role, preference alignment, risk, and evidence confidence. Debug details remain available to the system, while applicant-facing explanations are intended to surface the highest-value reasons rather than raw scoring output.

Current models can use specialty, applicant pathway, exam performance and attempts, visa context, research, clinical experience, letters, graduation timing, phase, CV evidence, and preferences when those fields are documented. Factors that are absent, ambiguous, or not represented in the model cannot be safely assumed. Programs may also consider school context, grades, AOA, away rotations, personal connections, and holistic factors that ResidencyLens cannot fully observe.

Different tools use different sources, assumptions, and objectives. ResidencyLens is designed to show its evidence and distinguish feasibility from preference and opportunity. Treat disagreements as a prompt to inspect the inputs, sources, and tradeoffs rather than as proof that one opaque ranking is correct.

No. The deterministic engines establish the core profile and strategy outputs. Ava explains them, answers questions, and may propose review changes, but an agent recommendation should identify a genuine difference from the current list and requires your approval before changing saved strategy.

The program universe combines local snapshots and public information from sources such as NRMP, AAMC ERAS, ACGME, and available public directory data. Specialty-specific public and community datasets may add context. ResidencyLens keeps official evidence, applicant-reported signals, and model inference as separate evidence types.

No. Much of the universe is built from dated local snapshots, not a live official feed. Each cycle can change participation, tracks, signals, visa rules, screening policies, and program details. Verify time-sensitive facts with the program and the relevant official service before acting.

Community evidence is treated as a lower-authority signal, not an official fact. Small samples, selection bias, anonymous reports, duplicated entries, and changing program culture can all distort it. Community data should influence interpretation only when provenance and support are strong enough.

ResidencyLens does not infer a visa, graduation-year, attempt, or score-screening policy simply because a value is absent. Unknown should remain unknown until a sufficiently authoritative and current source supports the field.

Use the program's official website and current ERAS, NRMP, ACGME, or ECFMG guidance as the deciding source. A local directory error or stale snapshot should never override current official information.

Applications build portfolio breadth across realistic opportunities. Signals are scarce and should express interest where the applicant-program fit, preference, opportunity, and likely value of signaling align. A good application option is not automatically the best use of a signal.

No preference means you are open geographically. It should not produce an empty shortlist or act as a hidden exclusion. The engine can still prioritize signal candidates using applicant fit, opportunity, and other stated preferences.

No. Programs interpret signals differently, and specialty rules can change by cycle. A signal recommendation is an allocation aid, not a promise of an interview or a quantified return.

Geography can shape signal strategy and preference ordering, but it should not erase applicant-program feasibility. ResidencyLens is designed to apply preferences on top of the fit model rather than turning every program in a selected region into a strong fit.

No. Use it as a structured second opinion. Personal circumstances, mentors, couples matching, family support, away rotations, home-program relationships, and career goals may justify choices that the model cannot observe.

The current beta supports Diagnostic Radiology, Internal Medicine, and General Surgery. Additional specialties should be added only after their program data, applicant calibration, signal rules, and evaluation cases are ready.

The current launch is a free beta and billing is inactive. If pricing or usage limits are introduced later, they should be disclosed before a user incurs a charge.

No. ResidencyLens is an independent decision-support product. It is not affiliated with or endorsed by NRMP, AAMC, ERAS, ACGME, ECFMG, individual residency programs, or your medical school.

Verify current program participation, track structure, eligibility, visa support, application requirements, signal rules, deadlines, and rank-list procedures through official sources. Your final choices remain yours.