Membership
Five ways to be tested.
The price is on this page, in full. So is the thing it protects. The tiers are not a ladder of features. They are a ladder of what is at stake.
Every membership runs the same thing underneath: a fleet of AI agents that audits your accounts, devices, and home network every night, on private local models. What rises with the price is not the software. It is the number of real people who test you, and how fast they answer when it counts.
What a membership buys
Testing, not insurance
Identity-theft products reimburse you after a loss. Feint tries the attack first, with your written permission, so the loss never happens. The fleet finds the openings; people decide which ones matter.
How the tiers differ
Depth, then people
Higher tiers run deeper audits and add frontier models. The real change is human: more exercises a year, a named lead who knows you, and a faster line when something is wrong.
How to read the price
Measured against the loss
Do not weigh the price against what it costs us to run. Weigh it against what one unprevented morning would cost you. The average loss for someone over 60 is $38,500. One prevented call pays for years.
Be honest about the overlap
Some of this you can get cheaper.
The lower tiers sit next to tools that now cost little or nothing. We would rather you heard that from us.
What cheaper tools already do
- Passkeys close most password attacks, and they are spreading. 90% of consumers know them; 75% have turned one on. FIDO Alliance, State of Passkeys 2026
- Phones are adding free scam detection. Google now screens Android calls and messages with on-device AI. BleepingComputer, 2026-10-15
- A year of phishing training cuts click rates from about 33% to 4%. Good training is cheap. KnowBe4, 2025-05-13
- For most people the typical scam loss is small. The median reported loss was $500. Stop Scams Alliance / Gallup, 2026-06-30
What Feint adds on top
- It tests your setup, not the average one. Passkeys help if they are on the accounts that matter. We check whether yours are, every night.
- It closes the gap between knowing and doing. Regular use of two-factor fell from 94% to 53% in three years. People know the advice and stop following it. National Cybersecurity Alliance, 2026-02-25
- It rehearses the rare, expensive call. The median loss is small, but six-figure losses are what grow fastest. One prevented call pays for years.
- It is one party holding the whole picture. Phone, bank, and training each guard their own corner. No one else tests the seams between them.
Where the money goes
Compute barely moves. Human attention climbs a cliff.
From Guard to The Second the monthly price multiplies about 3,165 times. The compute behind it multiplies about 257 times. The cost of people multiplies about 7,270 times. That gap is the product.
See the figures as a table
| Tier | Price / mo | Compute | People | Compute as % of price | Human hours / mo |
|---|---|---|---|---|---|
| Guard | $79 | $11 | $12 | 14.5% | 0.1 |
| Parry | $499 | $38 | $120 | 7.6% | 1 |
| Riposte | $5,000 | $129 | $1,700 | 2.6% | 9 |
| Master | $30,000 | $504 | $8,980 | 1.7% | 46 |
| The Second | $250,000 | $2,822 | $87,200 | 1.1% | 400 |
Figures from our cost model, model/pricing_model.py, 2026-10-21. Token counts are estimates until measured. Even switching every run to a frontier model would not change any tier's price.
I
Guard
The fencer's rest position. Blade up, watching.
$79/ month
$948 a year. No onboarding fee.
Who it is for
One person who wants a real check, not another subscription they forget. The unprepared majority: 73% of US adults have already had a scam or attack reach them.
What the fleet does
The agent fleet runs a small delta audit every day and a full sweep each month, on private local models. A delta looks at what changed: a new device, a leaked password, a setting that drifted. You see plain findings, never raw agent output. It is built to prove, in the first week, that it can find something real on your own accounts.
- Covers
- 1 person
- Human bouts / year
- None. This tier is software. Humans review what the fleet escalates.
- Response
- Your named human
- None at this tier
- Modeled human hours / mo
- 0.1
II
Parry
A defense that deflects the blade and sets up your answer.
$499/ month
$5,988 a year. Onboarding $500.
Who it is for
The security-curious professional, founder, or engineer. Especially the person who shipped an app with AI tools and now holds real users' data with no security team.
What the fleet does
Nightly local runs plus a monthly deep run on Sonnet. An analyst reads the output and turns it into a short, ranked fix list in plain language. You can send yourself phishing drills through standard vendor tooling, to practice on your own inbox. Four scans in a year found thousands of AI-built apps leaking personal data; this tier checks whether yours is one of them.
- Covers
- 1 person
- Human bouts / year
- Self-serve email drills. No human-led exercise yet.
- Response
- Email, within 2 business days
- Your named human
- None named; an analyst reviews every confirmed finding
- Modeled human hours / mo
- 1
III
Riposte
The counter-attack that follows a parry. The first tier where people test people.
$5,000/ month
$60,000 a year. Onboarding $7,500.
Who it is for
A whole household: parents, grandparents, teenagers, sometimes staff. This is the elder-fraud tier. Adults over 60 reported $7.75 billion in losses to the FBI in 2025, up 59%, with an average loss of $38,500.
What the fleet does
Nightly local runs and a monthly Opus deep run. First we help the family set a verification protocol: a passphrase, call-back numbers, the rule that money moves need two people. Then, once a quarter, we test it on one channel, with everyone's written consent. A debrief teaches; it never shames. Data-broker removal and a quarterly dossier show what an attacker would learn before touching anyone.
- Covers
- Household, up to 6 adults
- Human bouts / year
- 4 — one consented exercise each quarter, on one channel
- Response
- Hotline, business hours
- Your named human
- A shared lead who knows the household
- Modeled human hours / mo
- 9
IV
Master
The fencing master who runs the salle and drills you until the response is yours.
$30,000/ month
$360,000 a year. Onboarding $40,000.
Who it is for
A person who is a target by position: a founder, an executive, a public figure, a high-net-worth family. The attempt rarely comes at them directly. For Marcus, the real target was his assistant's inbox.
What the fleet does
Nightly deep local runs, a weekly Opus run, and a monthly Fable run on the hardest surfaces. Human exercises run every month, across email, text, voice, and in person, each inside a signed consent envelope. Quarterly on-site testing covers the home and office network. A monthly dossier tracks the family's public footprint. A single corporate red-team engagement runs $40,000 to $100,000 and more; this tier delivers that depth every quarter, aimed at one person.
- Covers
- Principal, household, and 2 staff
- Human bouts / year
- 12 — a monthly exercise, across all channels
- Response
- 24/7, with a 1-hour callback
- Your named human
- A dedicated lead
- Modeled human hours / mo
- 46
V
The Second
In a duel, your second stands beside you and handles everything.
By introduction
From $250,000 a month. $3,000,000 a year. Onboarding $150,000. Sold by referral; a few families a year.
Who it is for
One family, and everyone around it: the principal, relatives, staff, and the family office. 43% of family offices were attacked in the last two years. Close to a third had no plan for when it happened.
What you are given
A dedicated cell of three to five people and an agent fleet, working only on your family. They behave like a patient, well-funded adversary, entirely inside a signed consent envelope. Twice-daily local runs, a daily Opus run, and a twice-weekly Fable run. Unannounced exercises, physical and travel testing, and deepfake-readiness drills, each with the consent it needs. Vendor-chain review, with the vendors' own authorization. On-call forensics and legal partners. Most of the price buys scarce senior people and accepted liability, not tokens.
Why it is introduced, not sold
Capacity is the constraint, and so is trust. A cell holds a family's whole picture. We take on only families we can serve completely, by referral, after we have both decided it is a fit.
- Covers
- Principal, family, staff, and family office
- Human bouts / year
- Year-round. Unannounced, continuous, within the envelope
- Response
- 24/7, 15-minute callback, on-site anywhere
- Your named human
- A lead plus a cell of 3 to 5
- Modeled human hours / mo
- 400
Side by side
The five, compared.
The same rows, across every tier, for the reader who wants it all on one screen.
| Row | Guard$79 / MO | Parry$499 / MO | Riposte$5,000 / MO | Master$30,000 / MO | The SecondFROM $250,000 / MO |
|---|---|---|---|---|---|
| Per year | $948 | $5,988 | $60,000 | $360,000 | $3,000,000 |
| Onboarding | None | $500 | $7,500 | $40,000 | $150,000 |
| Covers | 1 person | 1 person | Household, up to 6 adults | Principal, household, 2 staff | Principal, family, staff, family office |
| Fleet cadence | Daily delta + monthly sweep, local | Nightly local + monthly Sonnet | Nightly local + monthly Opus | Nightly deep + weekly Opus + monthly Fable | Twice-daily + daily Opus + twice-weekly Fable |
| Human bouts / yr | None | Self-serve drills | 4 (one channel) | 12 (all channels) | Year-round, unannounced |
| Response | Email, 2 days | Hotline, business hours | 24/7, 1-hr callback | 24/7, 15-min callback, on-site | |
| Named lead | None | None | Shared lead | Dedicated lead | Lead + cell of 3–5 |
| Human hrs / mo | 0.1 | 1 | 9 | 46 | 400 |
Add-ons, when a family needs them
- Extra adult
- $250 to $2,500 a month, by tier
- Family-office scope
- From $10,000 a month, once the office signs its own authorization
- One-off deep audit
- $15,000 to $60,000 for a codebase or company
- Travel package
- $5,000 per trip: loaner devices, a briefing, a post-trip check
- Incident response outside tier
- $500 an hour, plus partner costs
The Second
Request an introduction.
The top tier is not bought from a page. Tell us a little, and if it looks like a fit, a partner will reach out to begin a private conversation. There is no obligation, and nothing is tested until everyone has signed.
This form is inert. It confirms on this page and sends nothing. Everyone pictured on this site is fictional; stories are illustrative composites.
Thank you. If this were live, a partner would be in touch within two business days. For now, nothing has been sent or stored. You might read how we work next.