Community Awareness for Invisible-Illness Reference and Navigation
An early-Lyme read that avoids a false all-clear.
Welcome to CAIRN
In the first weeks of Lyme disease, the standard blood test is frequently negative, because your body has not yet produced antibodies. Patients are sometimes sent home reassured before the infection can show on a test. Caught this early, a short course of antibiotics is often sufficient to clear it. This early diagnostic window is what CAIRN is built for.
CAIRN gives you an early-Lyme read. You answer a few quick questions, and it checks how they line up with the published guidance from the Centers for Disease Control and Prevention (CDC) and the Infectious Diseases Society of America (IDSA). It also shows how much Lyme is active near you, drawn from public federal data, with a clear range. You get a plain-language result and a one-page summary for your records. It is not a diagnosis, it helps you see whether your symptoms are worth a closer look.
CAIRN is free and runs on your device. What you enter stays on your phone unless you choose to share it, and your read holds no personally identifiable information (PII). There is no account, no sign-in, and no cost.
Tap your state to begin
Loading the CDC map…
Step one
Lyme activity varies by location and time of year. Provide your location to establish the seasonal baseline. This read evaluates your exposure window over the last three weeks, the recent window.
Please choose your state to continue.
Step two
Please answer the required questions above to continue.
Good to know
A little prevention goes a long way. How to lower your risk, what else a tick can carry, and what to do if you find one.
Lyme disease comes from a bacterium that blacklegged ticks, often called deer ticks, pass to a person through a bite. It is the most common tick-borne illness in the United States, with approximately 476,000 Americans diagnosed and treated each year (CDC). It concentrates in the Northeast, mid-Atlantic, upper Midwest, and parts of the West Coast.
Tick activity surges from late spring into summer, when young ticks are small and easy to miss. Caught early, Lyme disease is often highly treatable. The challenge is that the first signs, an expanding rash or a run of flu-like days, are easily shrugged off. That early diagnostic window is what CAIRN is built for.
Want to go deeper? The Johns Hopkins Lyme Disease Research Center maintains a clear, trustworthy guide.
Cover up. Wear long sleeves, tuck pants into socks, and closed shoes.
Repel. Use an EPA-registered repellent, and treat gear with permethrin.
Stay clear. Walk the center of trails, and avoid tall grass and leaf litter.
Inspect. Check yourself, your children, and pets after time outside, head to toe.
Wash and dry. Shower within a couple of hours, and tumble-dry your clothes on high heat to kill any ticks that caught a ride.
One tick bite can transmit more than one pathogen. In many regions, the same blacklegged ticks carry other infections that mimic Lyme but need different tests and treatment. Tap a card to see what each one is.
What to doShare your tick bite history with a clinician even if an early Lyme test returns negative, and ask whether testing for co-infections is appropriate.
Tweeze. Use fine-tipped tweezers. Grasp the tick close to the skin and pull straight up with steady, even pressure.
Clean. Wash the bite area and your hands afterward. Skip home remedies like heat, nail polish, or petroleum jelly, which irritate the tick and complicate removal.
Monitor. Note the date. Watch the spot and track how you feel over the next few weeks. Return to CAIRN to check your signs if symptoms develop.
Lyme tests look for your body’s antibodies, which take a few weeks to build, so a test taken in the first days or weeks can read negative even when Lyme is present. The standard approach is two-step antibody testing with FDA-cleared tests, ordered and read by a clinician. CAIRN does not run or interpret tests. If your signs or exposure point to Lyme, a negative early test is a reason to keep watching and follow up, not an all-clear.
Learn more at the CDCUnder the hood
Public, dated, and linked. CAIRN reads only open federal data for your county, and shows how current each source is and where it falls back.
CAIRN’s local estimate is built on reported Lyme cases from the Centers for Disease Control and Prevention (CDC), and that public data does not cover the country evenly. CAIRN handles the gap with a clear fallback, and shows it, rather than guessing.
We tested additional data types. For example, we evaluated terrain slope and, in states that track it, the local prevalence of Lyme-carrying ticks. Neither added a reliable signal for a national model, so CAIRN sets them aside. Where possible, we collaborate with researchers in pursuit of stronger data.
These sources are public and free. Explore U.S. open data at data.gov.
Stay ahead of ticksCommunity
CAIRN began close to home, with the people an early test most easily looks past. This is who we built it for, and why it is free.
“For years, my family suffered in silence with Lyme disease. Between my own delayed diagnosis and watching my two youngest children struggle with unexplained fatigue and brain fog, I would have given anything for a simple way to track our symptoms and show a doctor what we were going through.”
Mrs. Yang Retired teacher and mother of four Storrs, Connecticut
The hardest thing about early Lyme is how easily it hides. The standard test misses a large share of early cases, and it misses some people more than others.
Women, and especially premenopausal women, were less likely to test positive early, and more likely to have their symptoms read as something else.
Rebman AW, Yang T, Aucott JN. Sex and menopause-based differences in presentation of early Lyme disease. Clin Exp Med. 2026;26(1):139. doi:10.1007/s10238-026-02063-0
We take that seriously. One study is not yet enough to say by how much, or to safely change a person’s read based on their sex, so we handle that uncertainty with care. CAIRN avoids handing out a false all-clear. It weighs the early signs, and it treats an early negative test as a reason to keep watching.
As a physician, we are servants of the people, and as such, we need to listen to our patients in order to guide them to their maximum individual potential.”
Dr. Yasir J. Ahmad, MDBoard-certified psychiatristFounder of Hamilton Behavioral
Early Lyme is easy to miss, and an early test looks past some people more than others. CAIRN is our answer, a careful read that takes the early signs seriously and runs free on your device.
Get your readAbout us
A small engineering team building the Trust Layer for AI.
CAIRN is that work, put to the common good.
GREENBOX Labs is a small engineering team building the Trust Layer for AI: software that turns data into clear answers and stays transparent about its uncertainty. We work with federal programs and private companies on projects grounded in data. CAIRN is free, and built to collect no personally identifiable information (PII), because trust and integrity are core to what we do.
MLCommons memberNVIDIA Inception memberCAIRN began close to home. Lisa Yang, our Founder and CEO, watched her mother, sister, and brother all suffer with an unknown illness. Early Lyme is easy to miss, and the years it took for them to be treated were frightening. We built the reference tool we wished our family had during those critical first weeks.
Your result comes from a deterministic guideline tool built on published CDC and IDSA early-Lyme criteria. Consistent inputs yield consistent results, and no chatbot decides the medical read. A separate machine-learning spatial model, trained on public federal data, estimates local Lyme activity and gives a calibrated tier. That estimate gives context, and it does not change your clinical result. Each area number shows its public source.
NIST Trustworthy AI Aligned to global AI standardsCAIRN was built in the TOPx Tech Sprint at the U.S. Department of Health and Human Services, which put a spotlight on early Lyme disease and on the power of open public data. We are thankful for the opportunity to put our technology to work for the common good, showing how AI can draw meaningful insight from open federal data.
Terms and Privacy
Plain language, no surprises. CAIRN runs on your device, runs no ads, collects no personally identifiable information (PII), and does not sell your data.
Version 1.2, effective 19 October 2026, CAIRN by GREENBOX Labs
CAIRN, short for Community Awareness for Invisible-Illness Reference and Navigation, is a free tool from GREENBOX Labs, Inc. We built it for the public good: it reads public federal data for your area, on your device, and gives you a read of the early Lyme signs your answers match. It runs on your device, it runs no ads, and we do not sell your data.
CAIRN draws on public federal data, currently including the CDC, the U.S. Census Bureau, CDC ArboNET, the National Oceanic and Atmospheric Administration (NOAA), Gridded U.S. weather data (gridMET), the U.S. Geological Survey, the U.S. Forest Service, and HRSA, along with published CDC and IDSA guidance. This list is current, not exhaustive, and we may add, update, or remove sources at any time. Each source shows the vintage CAIRN’s model is built on, linked to the agency, and flags where a newer release exists. CAIRN is not affiliated with, or endorsed by, these agencies and organizations.
CAIRN provides general information, not medical advice and not a diagnosis. It does not constitute the practice of medicine. It does not diagnose or rule out any condition, does not create a doctor and patient relationship, and is not a substitute for professional medical judgment. Do not disregard professional medical advice, or delay seeking it, because of anything CAIRN shows you. If you think you may be having a medical emergency, call 911 or your local emergency number right away. Any decisions about your health are between you and a qualified healthcare provider of your choosing.
CAIRN is intended for people 13 and older. If you are under 18, please use it with a parent or guardian. CAIRN is not intended for children under 13, and we do not knowingly collect information from them.
CAIRN does not ask for your name, and it has no accounts or logins. Only if you opt in to share, we keep a de-identified read that holds your state, your county when you give a ZIP or county for the local picture, the month, your symptom categories (including any tick exposure, rash, and urgent signs), whether an early test came back negative, the result, and whether your local picture used a county, state, or left-out figure, along with any optional age band or sex you choose to share. A ZIP you enter is used only on your device to find your county; the ZIP itself is not kept. The record also carries a few technical details that are not about you, a random marker for your current browsing session, the time, a content seal for the read, and the app version, and these hold no personally identifiable information (PII). If you keep it to yourself, your read is not sent and stays on your device. When kept, this de-identified record is not linked to your identity. Like any website, CAIRN is served by hosting and service providers that may process technical data such as your IP address to deliver and protect the site. We do not attach that to your read, and we do not use it to identify you.
We use your de-identified read to operate and improve CAIRN and to support public understanding of Lyme disease. We do not sell, rent, or trade it, and we do not use it for advertising. We do not share it with other companies for their own advertising or AI training. Service providers that help us run CAIRN may process data only for that purpose, under contract.
Because the de-identified read is not linked to you, there is no account to delete. We keep it to run and improve CAIRN.
If you send feedback, we use what you share only to improve CAIRN and the public data behind it. Your email is optional. If you include it, we may reach out about your feedback. We do not use it for advertising, and we do not sell it.
When you finish a read, CAIRN asks one question: may we keep a de-identified copy of your read to help improve CAIRN, and keep optional usage analytics on this device. If you say yes, both are kept. If you keep it to yourself, your read is not sent and stays on your device, and no optional analytics are kept. The choice is real, and you can change it any time. Neither holds personally identifiable information (PII).
CAIRN is provided “as is.” It can be incomplete or wrong, and it is not a substitute for professional care. To the fullest extent permitted by law, GREENBOX Labs is not liable for any decision made in reliance on CAIRN.
We may update these terms; the effective date above will change when we do. These terms are governed by the laws of the State of Delaware, United States. Questions or requests: info@greenboxlabs.ai.
Model Card
A free, on-device educational tool for early Lyme disease.
It restates published CDC and IDSA guidance, not a diagnosis.
Part 1 of your read
What is a rule? A rule is a fixed set of if-this-then-that steps, like a checklist. There is no learning and no guessing, so it is fully transparent, and consistent answers give a consistent read. It is the right choice for the part that matters most to your health.
Your result maps your answers to published CDC and IDSA early-Lyme guidance.
Part 2 of your read
What is a model? A model is a formula built from real data. You give it things you can measure about a place, and it gives back an estimate of how much Lyme activity to expect. It learns the pattern from counties where the answer is already known, then applies that pattern to yours. That learning-from-data step is what people mean by “modeling.”
What the area model reads
The area estimate, tested on places it never saw
What real local tick data buys
Local environment data only
With real local tick surveillance added
A model gives an estimate, so it comes with a low-to-high range rather than a single certain number, and it adds context without changing your read. The two scores above come from counties it was never trained on. It is built and checked on real reported numbers:
For a technical reader. The area model is a ridge regression on log Lyme incidence, fitted once and frozen, so a given county consistently gets the same estimate. It is scored with state-grouped cross-validation, holding out whole states at a time across 10 folds, so each score comes from counties the model was not trained on. It is fitted on CDC reported cases for 2022–2023. It captures about 52% of the variation on the log scale, and 17% on the raw case-rate scale (R²). It reports a range rather than a single number, and that range is calibrated to contain the true rate about 9 times in 10.
On your device, CAIRN writes your read as an HL7 FHIR R4 record, with each reported sign mapped to a standard SNOMED CT code (SNOMED CT is maintained by SNOMED International) so a clinician or an electronic health record can read it directly. A sign with no standard code carries none, never a guessed one. The record holds no personally identifiable information (PII), and an automated check confirms it is reproducible, correctly coded, and PII-free across 80 input combinations.
HL7® and FHIR® are registered trademarks of Health Level Seven International. SNOMED CT® is a registered trademark of SNOMED International. CAIRN uses these standards to format the record and makes no claim of USCDI, US Core, or TEFCA conformance.
CAIRN reads only public, dated federal data. Every source, the year it is set to, and how current it is, live together on the Data page. When a county layer is not available, the state figure is used, and when neither is, the layer is left out rather than estimated.
Rigorous mathematical checks prove these safeguards hold for each read. They enumerate the app's reachable states and prove that a result is not shown without consent, that the area model does not change the medical read, and that a shared read carries only de-identified fields, with no personally identifiable information (PII).
CAIRN provides general information, not medical advice and not a diagnosis. It looks at early Lyme patterns only; other tick illnesses can be urgent even when this read is low. Surveillance data undercount real cases, and early tests can miss early infection, so a low read is not an all-clear.
The bigger picture
Caught in the first weeks, early Lyme is a short course of antibiotics. Missed, it becomes a bill, county by county across the States.
Tap a state to see its counties by tier.
A model we built
Each tier is a rule, not a guess. We take a county's two-year reported Lyme cases, average them to one year, multiply by the published per-case cost, and round to the dollar. The yearly total sets the tier, and the county figures are the record you can check. (Because each county is rounded, the national total is the exact sum of those figures, not a round multiple of the per-case cost.) CDC publishes county-level Lyme for 35 states, so those 1,958 counties carry a tier. The other 1,186, in 16 states with no county-level data, stay gray. There, a $0 means missing data, not a real low burden.
This is an area picture, not a diagnosis, and it does not change your read. The tier counts reported cases (which undercount true cases), applies one conservative per-case cost, and is figured at the county level.
