Vrymnox™ Systems LLC · Makers of Vrymdor™
A team of specialized AI agents watches your network, sorts real threats from noise, and prepares containment, using local AI models with zero cloud data egress. Anything high-risk waits for a person with a hardware security key. Available as Vrymdor™ Business Security and Vrymdor™ Home Security.
The Vrymdor™ film, about two and a half minutes. It plays muted with captions; turn on the sound with the controls.
The problem
Many AI tools are thin wrappers around a public cloud model. Your data leaves your network, automated actions can repeat with no one checking them, and the tool breaks when its vendor changes something.
How Vrymdor™ works
Each Vrymdor™ agent has one clear job and a defined set of steps it may take, running on hardware you control: in your cloud, on premises, or fully air-gapped. Telemetry only ever travels outbound.
The safeguard
Before an agent makes a risky change, such as deleting data or rewriting a firewall policy, a person must approve it by touching a physical security key. Nothing high-risk runs on its own.
Coming soon
Vrymdor™ Business Security and Vrymdor™ Home Security are coming soon. Official operations are in progress. Sovereign Vault sizing calls can be booked now.
Registered business
Vrymnox™ Systems LLC
Nebraska limited liability company
Public record with the Nebraska Secretary of State
Vrymnox™ Systems LLC · Products
Vrymnox™ Systems makes Vrymdor™ and sells it as two products. Both run their AI on hardware you control, keep your data out of the cloud, and make every high-risk action wait for a person with a hardware security key.
For organizations · Coming soon
A multi-agent autonomous SOC for your servers, domains, and IP ranges. Three plans: Core Guard, Active Defense, and Sovereign Vault, the dedicated-database plan. AI agents are included in every plan.
For households · Coming soon
The same local-first protection, built for a home network. It keeps watching when the internet line is cut, and your home's data stays in your home.
Works with your AI tools
Connect n8n, Make, Zapier, LangChain, and MCP clients through the Vrymdor™ AI Node. AI agents are included in every Business plan, and Vrymdor™ never charges per call. Zapier and Make bill their own usage separately, as they always do.
Vrymdor™ Business Security watches your network, sorts real threats from noise, and contains them with a team of AI agents that run on your own hardware. Zero cloud data egress. Local Ollama inference. A person with a hardware key approves anything high-risk.
A security operations center (SOC) watches a company's systems around the clock and responds to threats. Vrymdor™ does the repetitive first-line part of that work. Your site reaches out over one encrypted connection on port 443, so no inbound firewall port is ever opened, and onboarding takes one command.
The core platform is fully built. We're finalizing our digital Authorization Agreements; onboarding links go out to the waitlist as soon as they're ready.
Vrymdor™ Business Security
The waitlist has no limit. The first 20 to join get the founders rate: 20% off the list price of Core Guard or Active Defense, including extra servers and agents, for as long as their account stays active.
All 20 founder spots have been claimed. The waitlist is still open, and new members will start at the list price.
Vrymdor™ Business Security · 2 min 24 s
Pick a scene to jump straight to it. Installer, token and onboarding timings in the film are simulated.
Pricing
The first 20 to join the waitlist get the founders rate on Core Guard and Active Defense, including extra servers and agents, for as long as their account is active. 20 of 20 founder spots remain. Nothing is charged until onboarding opens and you choose a plan.
All 20 founder spots have been claimed, so new members start at the list prices below. Nothing is charged until onboarding opens and you choose a plan.
AI agents are included in every plan, with no per-call fees. Core Guard includes 3, Active Defense 10, and Sovereign Vault 25. Each extra active agent is $10 a month, up to 10 agents on Core Guard and 30 on Active Defense; Sovereign Vault has no agent cap.
Small teams protecting a handful of servers, sites and IP ranges
$239.20/ month
$299 list · founders rate for the first 20 on the waitlist
Includes 3 servers, then +$32 per server ($40), up to 9 servers.
Includes 3 AI agents, then +$8 per active agent ($10), up to 10 agents. No per-call fees.
$299/ month
Includes 3 servers, then +$40 per server, up to 9 servers.
Includes 3 AI agents, then +$10 per active agent, up to 10 agents. No per-call fees.
Growing companies that need vulnerability and phishing coverage
$559.20/ month
$699 list · founders rate for the first 20 on the waitlist
Includes 10 servers, then +$28 per server ($35), up to 24 servers.
Includes 10 AI agents, then +$8 per active agent ($10), up to 30 agents. No per-call fees.
$699/ month
Includes 10 servers, then +$35 per server, up to 24 servers.
Includes 10 AI agents, then +$10 per active agent, up to 30 agents. No per-call fees.
Regulated organizations that need a dedicated database
from$1,499/ month
Priced from a published formula. The founders rate does not apply to this plan.
Prices per month. HIPAA and PCI packs are optional.
A server is one server, domain, or IP range that Vrymdor™ watches. Prices are per month, in US dollars. The founders rate applies to Core Guard and Active Defense only, not to Sovereign Vault.
Vrymdor™ works with n8n, Make, Zapier, LangChain, and MCP clients through the Vrymdor™ AI Node. Vrymdor™ never charges per call; Zapier and Make bill their own usage separately.
Need more than 24 servers, a dedicated database, or HIPAA or PCI packs? Book a Sovereign Vault call.
| Compare plans | Core Guard | Active Defense | Sovereign Vault |
|---|---|---|---|
| Servers included | 3 | 10 | 25 |
| Each extra server, per month | $40, up to 9 | $35, up to 24 | $30, no cap |
| AI agents included | 3 | 10 | 25 |
| Each extra active AI agent, per month | $10, up to 10 | $10, up to 30 | $10, no cap |
| Per-call fees | None | None | None |
| Local AI (Ollama) triage and plain-English summaries | |||
| Perimeter and port discovery (Nmap and Naabu) | |||
| HTTP header and SSL/TLS monitoring | |||
| ntfy push alerts to your phone | |||
| Hardware-key operator controls, zero-egress storage | |||
| Vulnerability assessment (Nuclei) | Not included | ||
| Phishing simulations (Gophish) | Not included | ||
| Social engineering analysis and reporting | Not included | ||
| High-frequency scans and priority alert routing | Not included | ||
| Dedicated, isolated database | Not included | Not included | |
| HIPAA and PCI compliance packs ($400 a month each) | Not included | Not included | Optional |
| Log retention beyond one year ($150 a month per year) | Not included | Not included | Optional |
| Custom scan schedules and a direct operator channel | Not included | Not included | |
| Founders rate for the first 20 | Not included |
Sovereign Vault
Sovereign Vault is the plan with your own dedicated database. Because every organization needs a different amount of it, we start with a short call to learn what you need, then price it from the published formula.
01 · Recorded only with your consent
We ask before we record. If you agree, the call is transcribed offline on Vrymnox™'s own machine. The audio and the transcript are never sent to a cloud service or a third-party note taker.
02 · Your requirements profile
What you need is saved as your requirements profile: how many servers and AI agents you run, the compliance you must meet, and how long you must keep logs. We remember it, so you don't have to explain it again, and it stays with your workspace.
03 · A price from a fixed formula
The price comes from the formula on this page, not from guesswork: from $1,499 a month for a dedicated database with 25 servers and 25 AI agents, then $30 per extra server, $10 per extra agent, $400 for each HIPAA or PCI pack, and $150 a month for each year of log retention beyond one year.
04 · A workspace sized to you
Your workspace is set up from the same profile, so what you pay for is what you get.
Book a Sovereign Vault call
Sovereign Vault is the dedicated-database plan. On the call we learn what your company needs and price it from the published formula. Rough numbers are fine.
Inside Vrymdor™
Vrymdor™ is an enterprise-grade multi-agent autonomous SOC platform powered by the Vrymnox™ zero-egress architecture. Rather than relying on a single slow AI model, it runs a coordinated team of specialized, lightweight AI agents on your own hardware, each with one job. Select a part to see what it does.
The agents
The foundation
Agent 1 of 4
Continuously watches network and system telemetry from your firewall, switches and Wi-Fi, identity provider, endpoints, and Kubernetes audit logs, and flags activity that looks wrong.
In plain terms: the lookout
It reads the activity logs your equipment already produces, all day, every day, so nothing suspicious goes unseen.
Agent 2 of 4
Checks each alert against what Vrymdor™ already knows about your environment, kept in a local PostgreSQL database, to filter out false positives and rate how serious a finding is.
In plain terms: the analyst
It asks two questions about every alert: is this real, and does it matter? Most alerts are noise, so your team only sees the ones that are not.
Agent 3 of 4
Prepares a fast, local containment action, such as a firewall rule update or isolating a device from the network. Reversible steps like isolating a device can apply on their own and are undone with one click.
In plain terms: the responder
It contains a problem quickly, in a way that is easy to reverse.
Agent 4 of 4
Enforces a physical WebAuthn circuit breaker: any destructive or high-risk change to your systems waits until a person approves it with a hardware security key.
In plain terms: the safety lock
An AI agent can suggest a risky change, but only a person holding the key can make it happen.
Foundation 1 of 3
The agents use language models that run on your own server through Ollama. No alert, log, or finding is sent to a cloud AI service.
In plain terms: where the thinking happens
The AI does its work on your machine, not on someone else's.
Foundation 2 of 3
Every step an agent takes moves through defined states that are checked and stored in PostgreSQL, secured with SCRAM-SHA-256 authentication. Every decision lands in a tamper-evident record.
In plain terms: the playbook and the record
The agents follow a fixed playbook, and there is a permanent, checkable record of everything they did.
Foundation 3 of 3
Keeps the system from sending your data to cloud AI services: 0 KB of your logs leaves your network to be analyzed.
In plain terms: the locked door
Your data has no route out to outside AI companies.
Vrymdor™ is a coordinated system of agents with defined steps and checks, run by a deterministic state machine, not a chat model with a security label.
Connect n8n, Make, Zapier, LangChain, and MCP clients through the Vrymdor™ AI Node. Agents are included in every plan, with no per-call fees.
The agents take on repetitive first-line analyst tasks, so your people spend their time on decisions.
How it is built
No cloud telemetry and no cloud AI. Every scan, inference, and remediation decision runs on infrastructure you control, and only encrypted metadata travels outbound.
Operates entirely within your private network via WireGuard and local PostgreSQL. No third-party data harvesting or cloud relays.
Analyzes security events locally using Llama 3 via Ollama, never sending sensitive infrastructure logs to external cloud LLMs.
Complete fleet oversight and order tracking managed locally. No corporate cloud lock-in, no shared telemetry.
Scoped black-box and white-box assessments run on a schedule, surfacing exposed ports, outdated services, and misconfigurations as your stack changes.
Every scan, model inference, and finding stays on loopback. No cloud AI APIs and no usage telemetry; raw logs never leave your network.
Vrymdor™ Guardian runs open-source LLMs locally to translate findings into executive risk summaries, step-by-step code patches, and copy-paste vendor action plans.
Who Vrymdor™ is for
Whether you are a two-person startup or a public agency, Vrymdor™ scales with your servers, not your headcount. The same scan gives engineers a patch and owners a plain-English action plan, so it works whether or not you have a security team.
Not on the list? If you have domains or IP ranges to defend, Vrymdor™ is built for you too.
Build the evidence you need toward SOC 2 and close enterprise deals without hiring a $250k in-house security team. Continuous assessments, developer-ready patches, and an audit trail of what you fixed and when.
No security jargon. Get a plain-English picture of your risk and clear vendor action plans you can copy and paste straight to your IT provider, so the right things get fixed without you having to translate them.
What it covers
Nmap and Naabu map every open port and exposed service on your domains and IP ranges, so nothing is reachable without you knowing.
Nuclei checks the services you expose against known vulnerabilities and common misconfigurations.
Personalized Gophish campaigns train your own staff on your own domain. A wrong click leads to a short lesson, not a reprimand.
A local model ranks every finding and explains each security event in plain English. Nothing is sent to a cloud AI service.
Scheduled scans plus HTTP header and SSL/TLS checks catch new exposure as your stack changes, with push alerts to your phone.
Optional HIPAA and PCI packs turn on the matching controls and build exportable evidence from your scan history to support your audits.
Autonomous, reversible remediation
Every change Vrymdor™ proposes is sorted by risk before anything is touched.
Low risk: applied automatically
Low-risk fixes that can be undone and never touch production are applied for you. Each one is recorded in the audit trail, shown on your dashboard as auto applied, and emailed to you.
High risk: held for approval
Any production change, anything that can't be undone, and anything the policy can't classify waits as pending. It is applied only after strict human approval with a registered hardware security key (WebAuthn).
When in doubt, the policy fails safe: nothing is ever applied to production automatically.
Security awareness and education
Security works better when the people behind it understand what is happening and why.
Every event, explained
All plansEach security event on your dashboard comes with a short, plain-language explanation: what happened, why it matters, and whether you are safe. You don't need to be a security expert to follow it. The explanation is written by the local AI model, so nothing leaves for a cloud service, and the original record is never altered.
Phishing training, not a gotcha
Active Defense and upYour administrator runs simulations against your own staff, on your own verified domain, and never against outside addresses. The local AI writes a realistic, unique message for each person. If someone clicks, they land on a page that explains what they clicked and what to look for next time, followed by a 60-second review.
A staff member clicks a link in a simulated phishing email.
An education page explains the warning signs they missed.
They complete a 60-second interactive review, or get a friendly reminder to finish within 48 hours.
Only if it is still unfinished after 48 hours is their administrator or manager notified.
Names, emails, and departments used to personalize simulations are encrypted, stay in your tenant, are processed only by the local model, and cannot be exported.
Vrymdor™ Guardian is the local reasoning engine behind Vrymdor™, designed and built in-house by the founder. It runs open-source LLMs on your own hardware over loopback, keeps a vector memory of scan history and tech-stack profiles, and writes findings to encrypted storage, with no external cloud API ever in the path.
Loopback-only execution
All services bind to 127.0.0.1. Nothing is exposed to the public internet, by design.
Local open-source inference
Reasoning runs on your own hardware via Ollama (127.0.0.1:11434) using open models, with no third-party AI API calls.
Encrypted store + vector memory
Scan history, findings, and remediation are persisted to encrypted PostgreSQL with a local vector index.
Zero cloud telemetry
No usage analytics, no data egress. What the engine learns about your stack stays on your stack.
tls.header → HSTS missing on api.local
cwe-523 · port 443
synthesis → "Sessions can be downgraded…"
model: local · llama3
patch → Strict-Transport-Security header
copy-paste · nginx.conf
Onboarding Sandbox
Once official operations begin, six simple steps take you from joining the waitlist to your first security check-up. Each short animation shows what that step looks like, so you know exactly what to expect.
Fill in a short form: your work email, the website or network you'd like us to check, a plan, how many servers you need, and a card. It takes about two minutes, and your card is only charged if your account is approved.
A real person reviews every request by hand; nothing is approved automatically. Because we work across time zones, this can take a little while, but we'll get back to you as soon as possible. Once you're approved, the founder emails you to agree on a time for your onboarding call.
Before anything is tested, you sign a short digital agreement. It confirms that the website or network belongs to you, or that you have permission to test it, so everything stays legal and above board.
We check that the website or network you named really belongs to you, then lock it in as the only thing we're allowed to test. Nothing is scanned until this check is complete.
During your onboarding call, the founder helps you install Vrymnox™ on a computer or server you own. It runs entirely inside your network, so your data never leaves it.
Vrymnox™ runs its first check-up and explains each problem in plain English, along with how to fix it. Nothing changes without your say-so: you review every fix and approve it yourself.
Step 1 of 6
These screens are simplified illustrations: the names and data are made up, and nothing here is sent anywhere. Founding organizations also take part in a 100-day hardening and validation cycle, which starts when the cohort locks.
Start with step 1Founders Club
Anyone can join the waitlist. The first 20 to join help shape the product and keep the founders rate for life.
On Core Guard and Active Defense, including extra servers and agents, for as long as your account is active. Sovereign Vault is priced from its own formula and is not discounted.
Talk directly with the person who built Vrymdor™. Your feedback shapes what the product does next.
Launch-day setup and onboarding, handled personally by the founder.
[ 0 / 20 ] claimed
A founder spot is claimed when someone joins the waitlist. When the 20th spot is claimed, the founders rate closes and the waitlist stays open at list price.
All 20 founder spots have been claimed. Thank you to our founding members. The waitlist is still open at list price.
Testing protocol
100-day production-grade hardening and validation cycle for every founding organization, starting when onboarding opens.
01 · Now open
Waitlist
Join with your work email. There is no limit on the waitlist.
02 · First 20
Founders rate
The first 20 to join keep 20% off the list price of Core Guard or Active Defense for life. After the 20th, new members start at list price.
03 · Official operations
Onboarding opens
We email the waitlist in order. Every account is reviewed by hand, and nothing is charged before you choose a plan.
04 · Day 0 to Day 100
Production-grade hardening & validation
A 100-day cycle for founding organizations, run with the founder.
Only 20 founder spots. Join the waitlist to claim one.The waitlist is open. Join to hear first when onboarding opens.
Join the waitlistFAQ