Community
Try the AI Node on one or two agents, free.
$0/ month
Free. The included requests are a hard cap each month.
- Up to 2 AI agents
- 10,000 requests a month
- Hard cap at the limit
- Standard data masking
- Local audit log
- Default policy
Vrymdor™ AI Nodea privacy filter for AI
Use AI without giving away private data.
If you send customer emails, invoices or support tickets to AI, it hides the email addresses, card numbers and secret keys first, and keeps a record of every run.
- Small businesses
- Support teams
- Sales
- Finance
- HR
- Agencies
- Developers
- and many more
Free plan, no card needed.
The problem
AI reads everything you send it.
Every time an automation hands text to an AI model, the private details inside go along too. Unless something checks the text first, nothing stops them, and nothing records what was sent.
Support
Every new support ticket goes to an AI model for a summary. The customer’s email address and order details go with it.
Sales
Your AI assistant writes follow ups from a list of leads. Every lead on the list has an email address.
Finance
Invoices and receipts go to AI so it can pull out the totals. Card numbers can be printed right on them.
Sound familiar? That’s the gap the Vrymdor™ AI Node closes.
The solution
Meet the Vrymdor™ AI Node.
A privacy filter that sits between your apps and the AI model. It hides the private details first, so the AI gets your text without the email addresses, card numbers and secret keys in it.
Your appn8n, Make, Zapier, your code
Vrymdor™ AI Nodehides, checks, records
The AI modelgets clean text
New to the word? A node is one step in an automation. You add the AI Node as the step right before the one that talks to AI.
Who it’s for
More people than you think. Probably you.
It isn’t only for developers. If any tool you use sends customer or company text to an AI model, there are private details in that text. Now you can keep using AI and keep those details out of it.
Small business owners
An AI assistant answers your customer emails, and every message carries the customer’s email address.
Online stores
AI answers order questions and writes replies, and every order has a customer’s email address on it.
Customer support teams
Tickets go to AI for summaries and draft replies, often with email addresses and card details inside.
Sales and marketing
AI writes follow ups and sorts leads, and every lead is a person with an email address.
Finance and bookkeeping
AI reads invoices, receipts and statements to pull out the numbers, card numbers included.
Accountants and tax preparers
AI helps sort client files and tax forms, which carry Social Security numbers.
Recruiting and HR
AI sorts applications and new hire paperwork that carry email addresses and Social Security numbers.
Nonprofits
AI drafts thank you notes from donor lists and reads donation receipts, with emails and card numbers on them.
Agencies and freelancers
You build AI automations for clients, and their customers’ data runs through every one of them.
Operations teams
Order forms, contracts and chat transcripts get handed to AI to sort, summarize or answer.
Developers and IT
Your AI agents read logs, settings files and code, where API keys and tokens tend to hide.
Anyone building with AI
If you are wiring an AI model into a form, an inbox or a spreadsheet, real people’s details will pass through it.
What it hides today: email addresses, phone numbers, card numbers, US Social Security numbers (written like 123-45-6789), private keys, and API keys and tokens. It does not hide names or street addresses on its own; on Pro and Business you can add your own patterns for anything else.
Works with (setup for each)
- n8n
- Make
- Zapier
- GitHub Actions
- LangChain
- MCP clients
- Your own apps
If your automation talks to AI, put the Vrymdor™ AI Node in front of it.
Get started freeHow it works
Live in three steps.
No sales call, no approval, no setup project.
Step 1: Create your account
Sign up with your email. The free plan needs no card.
Step 2: Add your agent
An agent is the automation or AI assistant you want to protect. Give it a name, like “Sales Assistant”.
Step 3: Copy 3 values and you’re live
Paste them into the AI Node in your automation, and every run is filtered from then on.
Connect your AI Node
- Node Host URL
- https://your-node.example.com
- Agent ID
- agt_3f9c2a7d1b8e4f60
- Agent Secret
- ••••••••••••••••
What you get
Five things you get from the first run.
Private details stay out of AI
Email addresses, phone numbers, card numbers, US Social Security numbers, private keys, and the secret keys for OpenAI, Anthropic, Stripe, GitHub, AWS, Google and Slack are hidden before the model reads the text.
Your rules on every call
Your policy decides whether each call is allowed or blocked, and every block comes with a reason in plain English.
Proof of what happened
Every run writes an audit record: the decision, and a count of everything that was hidden.
Works with the tools you already use
n8n, with an official community node on npm, plus Make, Zapier through webhooks, GitHub Actions, LangChain and MCP clients.
Free to start
Community is free for up to 2 agents and 10,000 requests a month. Pro and Business start with 14 days free.
Who hosts what
Two questions, two different answers.
“Hosting” means who runs a piece of software on their computers. Your automation tool and your Vrymdor™ AI Node can each be hosted by someone else or by you, and the two don’t have to match.
1. Who runs your automation tool?
The tool’s company
n8n Cloud, Make, Zapier, and GitHub Actions on GitHub’s own runners
Your workflow runs in their cloud and you use it on their website. It can only reach an AI Node that has an internet address.
You
n8n on your own server, your own app, Claude Desktop, Cursor or VS Code on your computer, GitHub self hosted runners
Your workflow runs on a machine you control, so it can reach an AI Node on the same machine or network.
2. Who runs your Vrymdor™ AI Node?
Pro, Business and Enterprise
Vrymdor hosts it for you
- We run your AI Node on our own node server, a server Vrymdor operates in the cloud (rented from DigitalOcean, listed on our Subprocessors page).
- Your node is yours alone: its own container, storage, network and keys, separate from every other customer.
- Nothing to download or install. In your dashboard open Home base, Hosted node and press Turn on; in about a minute it shows your Node Host URL, which looks like https://node.vrymdor.com/t/k3v9q2m7x4ab.
- Your text passes through your node and is not stored. Your dashboard keeps only a log of what happened: the time, the agent, the decision and how many values were hidden.
- The AI provider keys you add are locked so only the node server can read them, and they are never shown again.
Every plan, and the only option on the free Community plan
You host it yourself
- You download the AI Node from your dashboard and run it with Docker on your own computer or server.
- Your text stays on that machine and in your network. The node connects to your dashboard to get your agents and rules and to report that it is running.
- Your Node Host URL is that machine’s address on port 8080.
- A tool in someone else’s cloud (n8n Cloud, Make, Zapier, GitHub’s runners) can reach it only if you give that machine an internet address with HTTPS, for example behind your own reverse proxy.
Which fits your setup
Your tool runs in its company’s cloud (n8n Cloud, Make, Zapier, GitHub’s runners)
Use: Let Vrymdor host your AI Node (Pro and up). On Community you can still do it, but you must run the node on a machine with a public HTTPS address.
Your tool runs on your own machine or network (self hosted n8n, your app, Claude Desktop, Cursor, VS Code)
Use: Either works. Running the node yourself next to the tool is the simplest on Community; Vrymdor hosting means nothing to install.
One run, start to finish
Your automation tool
Your workflow sends the text to your Node Host URL, signed in with your Agent ID and Agent Secret.
Your Vrymdor™ AI Node
It hides the private details and checks your rules, on Vrymdor’s node server or on your own machine, wherever your node runs.
The AI model
Only the cleaned text goes on to the model, using your own provider key, or the cleaned text goes straight back to your workflow.
On the way back
The answer is checked the same way before your workflow gets it.
Your dashboard
A record of the run: the time, the agent, the decision and how many values were hidden. Never the text itself.
Set up on your platform
Pick your tool. Start today.
Every tool takes the same three values from your dashboard: Node Host URL, Agent ID and Agent Secret.
Same value, different names: your dashboard shows the Node Host URL as “Your node address”.
The three values, exactly
Node Host URL
Looks like: https://node.vrymdor.com/t/ and 12 lowercase letters and digits (hosted by Vrymdor), or your machine’s address ending in :8080 (hosted by you)
https://node.vrymdor.com/t/k3v9q2m7x4ab
Dashboard: Home base, Hosted node (or your own machine’s address). A hosted node keeps its address when you turn it off and on; in the rare case it moves to a new server, Home base shows the new address, so copy it from there.
Agent ID
Looks like: agt_ and 16 lowercase letters and digits, 20 characters in all
agt_3f9c2a7d1b8e4f60
Dashboard: AI agents, after you register an agent
Agent Secret
Looks like: vrym_agent_ and 48 letters and digits, 59 characters in all
vrym_agent_ followed by 48 characters (shown only once)
Dashboard: AI agents, right after you register an agent
When you paste them
- Use the Copy button next to each value. It copies the exact value, with nothing missing and nothing extra.
- No spaces at the start, at the end or inside any value, and no quotation marks around them.
- No slash at the end of the Node Host URL. The tool adds the rest of the address itself.
- Letters must stay exactly as they are: the Agent Secret mixes capital and small letters.
- The Agent Secret is shown only once. If you lose it, register a new agent and copy the new pair.
Waiting for approval, and why it doesn’t stop you
- n8n Cloud: n8n is still reviewing the node for its verified list, and n8n Cloud only installs verified nodes. The workflow file does the same job today.
- Make: Make is still reviewing the app for its public list. Until then we share it with your Make organization directly.
- GitHub Actions: Our GitHub Marketplace listing is waiting for GitHub’s approval. It only adds a way to pay through GitHub; you don’t need it to use the AI Node on GitHub.
Every tool below works today.
Connect your tool
Not sure where your AI Node should run? See who hosts what first.
n8n (self hosted)Ready nowThe official community node, installed by name from npm.
- In n8n, open Settings, Community Nodes, Install, and enter n8n-nodes-vrymdor.
- Add the credential Vrymdor AI Node API and paste your three values.
- Add the Vrymdor AI Node node, choose Remove PII and Secrets, and press Execute step.
Shortcut: download the ready workflow and choose Import from File.
You’ll know it works whenSaving the credential tests it at once. The first run shows [REDACTED:EMAIL] where the test text had an email address.
If it doesn’tIf the credential test fails, copy the Agent ID and Agent Secret again with the Copy buttons, and check the Node Host URL is a node n8n can reach.
n8n CloudReady now, with a downloadA ready workflow file you import, built from n8n’s own steps.
n8n is still reviewing the node for its verified list, and n8n Cloud only installs verified nodes. The workflow file does the same job today.
- Download the n8n Cloud workflow and, in n8n, choose Import from File.
- In its Vrymdor token step, paste the full URL from your Vrymdor dashboard into URL. It already ends in /v1/token.
- In that step’s Custom Auth credential, paste your Agent ID and Agent Secret.
- Press Execute workflow. The result shows [REDACTED:EMAIL] where the sample had an email.
Download the n8n Cloud workflow
You’ll know it works whenExecute step on Vrymdor token shows an access_token. Then Execute workflow shows [REDACTED:EMAIL] and [REDACTED:CREDIT_CARD].
If it doesn’t404 means the URL is missing /v1/token or has the wrong node address: paste the full URL from your dashboard again. 401 or “Authorization failed” means the Agent ID or Agent Secret is wrong: copy both again. 502 means your node is off: turn it on in Home base and copy the URL fresh.
A node n8n Cloud can reach over the internet. The hosted node (Pro and up) is the simplest.
ZapierReady nowTwo Webhooks by Zapier steps.
- Add Webhooks by Zapier, event POST, to your Node Host URL followed by /v1/token, with your Agent ID and Agent Secret as data.
- Add Webhooks by Zapier, event Custom Request, POST to your Node Host URL followed by /v1/node/run, with the access token from step 1.
- Test it, then map the cleaned output into your next step.
You’ll know it works whenTesting the first step returns an access_token. Testing the second returns "decision": "allow" and the cleaned text.
If it doesn’tNo access_token means a wrong Agent ID or Agent Secret, or a Node Host URL Zapier cannot reach.
A node Zapier can reach over the internet. The hosted node (Pro and up) is the simplest.
MakeReady now, by invitationThe AI Node app for Make (listed in Make as Vrymdor AI Node), shared with your organization.
Make is still reviewing the app for its public list. Until then we share it with your Make organization directly.
- Email [email protected] with your Make organization name, and we invite it to the app.
- In a scenario, add the module Run a governed action and create a connection with your three values.
- Choose Remove PII and secrets and press Run once.
You’ll know it works whenRun once shows decision allow and the output with [REDACTED:EMAIL] in place of the email address.
If it doesn’t“Your AI Node did not accept this Agent ID and Agent Secret” means one of them is wrong: copy both again.
A node Make can reach over the internet. The hosted node (Pro and up) is the simplest.
GitHub ActionsReady nowA workflow file and three repository secrets. Your dashboard walks you through it.
Our GitHub Marketplace listing is waiting for GitHub’s approval. It only adds a way to pay through GitHub; you don’t need it to use the AI Node on GitHub.
- In your dashboard, open Protect a GitHub repository and press Create a GitHub agent.
- In the repository on GitHub, add the three values as repository secrets (Settings, Secrets and variables, Actions).
- Add the workflow file the dashboard gives you as .github/workflows/vrymdor-node.yml and commit it.
- Open a test pull request: the sanitize check passes. Add a file named .env and it fails with Key file in the pull request.
You’ll know it works whenOn a pull request that only says what it changes, the sanitize check passes. Add a made up email address to its description and the check fails with Personal data in the description.
If it doesn’tA failed token request means a secret name or value is wrong: the names must be exactly VRYMDOR_NODE_URL, VRYMDOR_AGENT_ID and VRYMDOR_AGENT_SECRET.
The dashboard’s GitHub setup uses the hosted node (Pro and up). On the free plan, the guide shows the same workflow with a node your runners can reach.
Claude Desktop, Cursor and VS CodeReady nowMCP (the way these apps connect to tools), through a small bridge you run with uvx. No Docker needed, and it works with antivirus HTTPS scanning such as Norton.
- Install uv once (one command from the guide). It brings its own Python.
- Claude Desktop: paste the one command from AI agents, Connect your AI Node, MCP. It tests your agent, writes Claude’s settings and restarts Claude. Cursor and VS Code: add the vrymdor server to the MCP settings file.
- Click + next to the message box, then Connectors: vrymdor is listed. Ask Claude to run vrymdor_status, then to use vrymnox_sanitize on a test sentence.
You’ll know it works whenIn Claude Desktop, + next to the message box, then Connectors, lists vrymdor. Ask Claude "run vrymdor_status": it answers Connected, with your node and agent. vrymnox_sanitize answers "mail [REDACTED:EMAIL]" for "mail [email protected]".
If it doesn’tvrymdor is listed but says Not connected: its vrymdor_status tool says exactly what to fix. Not listed at all: run the one command again, and it says what is wrong. If an app cannot find uvx, put its full path in command (where.exe uvx on Windows, which uvx on macOS and Linux).
Python, LangChain and LlamaIndexReady nowThe Python client from PyPI: pip install vrymdor-node.
- Install it with pip install vrymdor-node.
- Create VrymnoxNode with your three values and call node.sanitize(...).
You’ll know it works whennode.sanitize("Customer [email protected] called") returns "Customer [REDACTED:EMAIL] called".
If it doesn’tAn unauthenticated error means a wrong Agent ID or Agent Secret: copy both again.
Your own appReady nowTwo web requests, from any language.
- Send your Agent ID and Agent Secret to your Node Host URL followed by /v1/token to get an access token.
- Send your text to /v1/node/run with that token. The answer holds the cleaned text.
You’ll know it works whenThe first request returns an access_token; the second returns the cleaned text in output.
If it doesn’tA 401 answer means a wrong Agent ID or Agent Secret, or a token older than 10 minutes: get a new token.
What the downloads are
The node (Download the node)
The AI Node itself, as a Docker image file for amd64 (most servers and PCs) or arm64 (ARM servers and Apple silicon), with a checksum file to check it.
- Who needs it
- The free plan, or anyone who wants the node on their own machine. Not needed with the hosted node.
- Where
- Your dashboard: Home base, Connect your node, Download the node. The link lasts 15 minutes.
Ready n8n workflows
n8n workflow files with the AI Node step already in them and a note with the steps.
- Who needs it
- n8n users who want the fastest start, and every n8n Cloud user.
- Where
- On this site: the self hosted workflow and the n8n Cloud workflow.
The n8n community node needs no download: n8n installs it by name from npm. Stuck on any step? Email [email protected] and a person answers.
Pricing
Start free. Upgrade when you grow.
Sign up yourself and start free. Community is free for good, with a hard monthly cap so nothing is ever charged. Pro and Business start with 14 days free: your card is saved at checkout and first charged when the 14 days end. On Pro and up we can host the node for you.
Pro
A growing team that wants its own policies and exports.
$29/ month
14 days free, then per month, plus $0.001 per request over the included amount.
- Up to 15 AI agents
- 100,000 requests a month
- Overage billed, never blocked
- Everything in Community
- Custom masking patterns
- Response filtering
- Log export
- Priority routing
Business
A larger fleet that needs roles and policy as code.
$99/ month
14 days free, then per month, plus $0.0005 per request over the included amount.
- Up to 50 AI agents
- 500,000 requests a month
- Overage billed, never blocked
- Everything in Pro
- Custom policy as code
- Role-based access control
- Advanced telemetry
Enterprise
Air-gapped or USB delivery, sized to you.
Custom
A contract sized and priced for your fleet, air-gapped or on a USB kit.
- AI agents sized to your contract
- No metered request limit
- Air-gapped or USB delivery
- Everything in Business
- Dedicated onboarding
- A signed Order Form, priced to your fleet
Requests are the calls your agents make through the node each month. Community is a hard cap: at its limit requests are blocked and nothing is billed. A paid tier bills only the overage and is never cut off. Prices are per month, in US dollars. After you sign up, the step by step setup guide connects your node to your platforms.
| What each tier includes | Community | Pro | Business | Enterprise |
|---|---|---|---|---|
| Where the node runs | You run it (Docker, GitHub Actions, your server) | We host it for you, or you run it | We host it for you, or you run it | We host it for you, or you run it |
| Live timeline of every agent action in the Vrymdor™ app | Not included | Included (hosted) | Included (hosted) | Included (hosted) |
| Plain English reason for every block | Included | Included | Included | Included |
| AI agents | Up to 2 | Up to 15 | Up to 50 | Sized to your contract |
| Requests each month | 10,000 | 100,000 | 500,000 | No metered limit |
| Past the included requests | Blocked until next month | $0.001 each | $0.0005 each | Not metered |
| Standard data masking, local audit, default policy | Included | Included | Included | Included |
| Custom masking patterns, response filtering, priority routing | Not included | Included | Included | Included |
| Log export and evidence pack (from a node you run) | Not included | Included | Included | Included |
| Advanced policy (gRPC rules, relaxing the safe defaults) and role-based access control | Not included | Not included | Included | Included |
| Advanced telemetry | Not included | Not included | Included | Included |
| Air-gapped or USB delivery, dedicated onboarding | Not included | Not included | Not included | Included |
| Sign up | Self-serve, free | Self-serve, 14 days free | Self-serve, 14 days free | By email |
Trust
Where your data goes, in plain words.
Where your data goes
Real values stay in your AI Node's memory only. They're never saved, logged or sent to the AI model or to Vrymdor, and they're deleted after 15 minutes by default (1 hour at most). Secrets are never put back. If anything fails, the request is blocked, never sent unchecked.
Works with
n8n, with the Vrymdor™ AI Node community node, and any app over the API.
API
curl -s -X POST "$NODE_HOST_URL/v1/node/run" \
-H "Authorization: Bearer $TOKEN" -H 'Content-Type: application/json' \
-d '{"action":{"type":"sanitize"},"input":"Mail [email protected]"}'
# "output": "Mail [REDACTED:EMAIL]"n8n
Vrymdor AI Node step (n8n-nodes-vrymdor)
Operation: Call a Model
Rehydrate: on
Only For: {{ $json.customer_email }}FAQ
Questions, answered plainly.
What is an AI Node?
A node is one step in an automation, like a step in an n8n or Make workflow. The Vrymdor™ AI Node is the privacy step: you put it in front of the step that sends text to an AI model, and it hides the private details in that text first. The AI model still gets the rest of the text, so it can do its job.
What does it hide?
Email addresses, phone numbers, card numbers, US Social Security numbers, private keys, and the secret keys and tokens for OpenAI, Anthropic, Stripe, GitHub, AWS, Google and Slack, plus sign-in tokens (bearer tokens and JWTs). Each one is replaced with a label such as [REDACTED:EMAIL], so the AI model still knows something was there. Secrets are also found when they are hidden in base64, hex, ROT13, reversed text, spaced out letters, look-alike characters or invisible characters, even in layers up to 3 deep (such as base64 of hex), and they are cleaned from the AI's replies too. It does not hide names or street addresses on its own; on Pro and Business you can add your own patterns for anything else.
Can my customer get the real details back in the AI's answer?
Yes, with Rehydrate. Each private detail becomes its own numbered label, such as [REDACTED:EMAIL_1], so the AI model can write its answer using the labels. Your own AI Node then puts the real details back in place of the labels before the answer reaches your customer. The AI model never sees the real details at any point. The list of labels and real details stays only in the memory of your own AI Node (on your server, or in your own container when Vrymdor hosts your node), for the time you set (15 minutes by default, one hour at most), and is then deleted. It is never sent to the AI model, never stored in Vrymdor's dashboard or records, and never written to disk, logs or backups. Restore is all or nothing: if the AI makes up or changes a label, nothing is put back. Secrets such as API keys are never put back. Your app can name who the answer is for (for example their email address), and then only that person's details can come back. Rehydrate is optional, and you can turn it off in your rules.
Can secrets ever be restored?
No. API keys, passwords, tokens, private keys and your own secret patterns are never put back, on any plan and with any setting. A request can't ask for them, they can't be put on the restore list, and the list of real values never leaves your AI Node. Secrets are also hidden in the model's replies, even when reply cleaning is off.
What is 'only for'?
A setting for putting the real values back. Your app names who the answer is for by their real details, such as [email protected], and only that person's values can come back. If the answer holds anyone else's details, or the person you named isn't in the message, nothing is put back. Use one person per request. In n8n it is the Only For option on Call a Model with Rehydrate on, or on Put Real Values Back; over the API it is only_for. Your rules can require it on every restore, which we recommend for anything sent to customers.
What happens during an outage?
It fails closed. If anything fails or times out, the request is blocked, never sent on unchecked. In n8n, Safe Mode then masks the text locally inside your workflow, so nothing goes out unmasked. Real values can't be put back during an outage, because they only live in your AI Node: the answer goes on with its tokens.
What doesn't Vrymdor protect?
We say it plainly: names of people, street addresses and dates of birth; bank account and routing numbers, IBANs, and IDs from outside the US; private information described in plain words (the node matches patterns, it doesn't understand meaning); a custom pattern that's written too narrowly; a correct token put back in the wrong place when 'only for' isn't used; prompt injection beyond known phrases (the check is one warning signal, not a guarantee); and anything sent to a model without going through the node. Personal data (emails, card numbers, US Social Security numbers, phone numbers) hidden in base64, hex or other encodings is usually caught too, but we don't promise it.
Is the Arena real?
Yes. The Vrymdor™ Arena at app.vrymdor.com/arena runs real attacks against an isolated copy of the current AI Node, never the copy our customers use and never customer data. Every value the Rivals use is fake, every run records the node version it ran on, and plain, fixed code decides whether anything got through, never an AI model.
What happens to my idea?
You send it from the Arena with a fake key we give you. We review it and approve it for the next Rival Run, merge it with an earlier idea that is the same (the credit stays with whoever sent it first), or reject it with a reason. After the run we email you the result. If it was blocked, you earn the participant badge. If it got through, we fix the AI Node, the attempt becomes a test every later version must pass, and the patch log credits your public name, or "a community member" if you post anonymously.
Do I need to be technical?
You need an automation tool you already use, such as n8n, Make or Zapier, and to be comfortable adding a step to it. You create an account, add your agent, and copy three values (Node Host URL, Agent ID and Agent Secret) into the AI Node. The step by step setup guide shows every screen.
How long does setup take?
There are three steps: create your account, add your agent, and copy three values into the AI Node. Most of it is copy and paste, and the free plan needs no card.
What does it cost?
Community is free: up to 2 agents and 10,000 requests a month, with a hard cap so nothing is ever billed. Pro is $29 a month for up to 15 agents and 100,000 requests. Business is $99 a month for up to 50 agents and 500,000 requests. Pro and Business start with 14 days free: your card is saved at checkout and first charged when the 14 days end. Enterprise is priced to your contract.
Who hosts the AI Node?
You choose. On Pro, Business and Enterprise, Vrymdor can host it for you: we run your node on our own node server, a server Vrymdor operates in the cloud (rented from DigitalOcean), with its own container, storage, network and keys for each customer, so there is nothing to install. On every plan, including the free Community plan, you can instead download the node and run it yourself with Docker, so your text stays in your own network. Your automation tool is separate: n8n Cloud, Make and Zapier are hosted by those companies, and n8n can also run on your own server.
Is my data safe?
The AI Node hides the private details before your text reaches the AI model, and it writes an audit record of every run, so you can see what was hidden and what was allowed. You can run the node yourself in your own network, or on Pro and up Vrymdor hosts it for you on our own node server; either way your text is not stored by Vrymdor, only a record of what happened. Your Agent Secret works like a password: keep it private, and never paste it into a chat, an email or an AI prompt.
What is a request?
One call your automation makes through the AI Node, for example one support ticket sent for a summary. On Community the node stops at 10,000 requests a month and nothing is billed. On Pro, requests past 100,000 cost $0.001 each, and on Business, requests past 500,000 cost $0.0005 each. A paid plan is never cut off.
Which tools does it work with?
n8n (with an official community node on npm), Make, Zapier through Webhooks by Zapier, GitHub Actions, LangChain, Model Context Protocol (MCP) clients such as Claude Desktop, and your own apps over HTTP. Zapier and Make bill their own usage separately.
Who makes the Vrymdor™ AI Node?
Vrymnox™ Systems LLC, founded by Ernest Hausmann. Vrymdor™ is our product name, and the AI Node is the one product we sell.
Start today
Keep using AI. Stop handing it private data.
Free for up to 2 agents and 10,000 requests a month. No card needed.