Anonymization as a node
The anymize engine is built in as a native block. A workflow can anonymize data before an LLM call and translate it back afterwards – PII never leaves your own infrastructure.
Workflows.
Automated.
Anonymized.
A visual no-code builder – with anonymization built in.
Workflows is a visual automation builder embedded directly into anymize. Drag building blocks onto the canvas, connect them into a flow and let a Rust executor run them. The decisive difference to n8n, Zapier or Make: you can anonymize text before an LLM call and translate it back afterwards. Personal data never leaves your infrastructure.
What are Workflows?
Automation
you can see.
A workflow is a chain of building blocks on an infinite canvas. You drag nodes onto the canvas, connect them with edges into a flow and configure each one through its inspector panel. No code, no YAML, no hidden logic – you see exactly what happens and when.
The canvas
01
Trigger
02
Anonymize PII
03
AI Agent
04
De-anon. PII
05
Output
You know the idea from n8n, Zapier or Make. But anymize goes further in two crucial ways:
Two extensions
Rust executor
Workflows are executed by a Rust-based executor – fast, deterministic and robust, with a step-by-step debugger and run history right inside the editor.
The canvas becomes your second workspace: on the left the chat you use in anymize every day – on the right the builder, where you create recurring flows once and never do them by hand again.
The AI copilot
Describe it.
The AI builds it.
You don't have to place a single node by hand. Describe the flow in one sentence – the copilot creates the trigger, AI agent and integrations live on the canvas, sensibly connected and pre-configured.
Four examples
Your sentence
"When a new Outlook email arrives, summarize it with an agent and send the summary to Teams."
Appears live on the canvas
Outlook trigger
AI Agent
Send to Teams
Your sentence
"Create a weekly revenue report from Google Sheets."
Appears live on the canvas
Schedule (cron)
Read Sheets
AI Agent
Report
Your sentence
"Automatically triage incoming support tickets."
Appears live on the canvas
Webhook trigger
AI Agent
If/Else
HubSpot
Your sentence
"Send the text anonymized to an LLM and turn the answer back."
Appears live on the canvas
Trigger
Anonymize PII
AI Agent
De-anon. PII
The copilot is not a rigid template system. It understands your context, picks the right integrations and connects the nodes in the correct order. You then refine in the inspector – or let it keep adjusting.
The building blocks
Everything a
flow needs.
Every workflow is assembled from nodes. Four groups, one inspector panel per node:
Triggers
The starting point of every workflow. Manually on click, via an external webhook, from the anymize chat, or on a schedule.
Manual
Webhook
Chat
Schedule (cron)
AI
The AI Agent is an LLM with a tool-use loop: it can autonomously research, calculate and act. Models: anymize Waterfall EU as well as Claude Opus, Sonnet and Haiku.
AI Agent
Anonymization & knowledge
The anymize-exclusive nodes – plus semantic search in your knowledge base. This is the decisive difference to any other builder.
Anonymize PIIanymize-exclusive
De-anonymize PIIanymize-exclusive
Knowledge-base search
Logic & connection
If/Else with expressions for branching, a generic HTTP request for any API, and native app integrations for your business applications.
If/Else
HTTP request
App integrations
The unique selling point
The only workflow builder
with anonymization built in.
n8n, Zapier and Make send your raw data unprotected to external LLMs. anymize does not. The anonymization engine is built in as a native node – a workflow can anonymize text before an LLM call and translate the result back afterwards. PII never leaves your own infrastructure.
The standard pattern
01TriggerReal data comes in
02Anonymize PIINames, addresses, IBAN → placeholders
03AI Agentonly sees [NAME_1]
04De-anonymize PIIPlaceholders → original values
05OutputResult with real data
What the AI agent sees
Your original
Dear Dr. Maria Schmidt, …
What goes to the LLM
Dear [NAME_1], …
The mapping
Encrypted
The assignment of placeholder to original value is stored encrypted – never in plaintext.
TTL 24 hours
The mapping lives at most 24 hours. After that, the back-translation is no longer possible.
Never in the logs
Neither original values nor the mapping appear in logs or the run history.
Integrations
Connected to
your stack.
Native app integrations for the tools you work with every day – plus a generic HTTP request for everything else.
Microsoft 365
- Outlook
- Calendar
- OneDrive
- Teams
- To Do
- OneNote
Google Workspace
- Gmail
- Calendar
- Sheets
- Drive
- Docs
HubSpot
- Contacts
- Deals
Notion
- Pages
- Databases
Asana
- Tasks
- Projects
GitHub
- Issues
- Pull requests
Postgres
- Query
- Insert / update
Email / SMTP
- Send
- Receive
Through the generic HTTP request node you reach any other API – without a new integration.
Tools for the AI agent
The AI agent works with a tool-use loop. It can call these tools autonomously:
Calculator
HTTP request
Knowledge base
Anonymize PIIexclusive
De-anonymize PIIexclusive
Microsoft / Google / HubSpot
Use cases
What workflows
actually solve.
Four flows you build once and never do by hand again:
Email triage
Typical flow
Outlook trigger → anonymize → AI agent classifies → de-anonymize → Teams
Apps
Outlook · Teams
Anonymization
Mail content is anonymized before the agent reads it
Report automation
Typical flow
Schedule → read Sheets → anonymize → AI agent summarizes → de-anonymize → PDF
Apps
Google Sheets · Drive
Anonymization
Numbers stay, personal names are anonymized
Ticket classification
Typical flow
Webhook → anonymize → AI agent → If/Else → HubSpot
Apps
Webhook · HubSpot
Anonymization
Customer data never leaves the infrastructure
CRM enrichment
Typical flow
Trigger → knowledge-base search → anonymize → AI agent → de-anonymize → HubSpot deal
Apps
HubSpot · Knowledge base
Anonymization
GDPR-compliant enrichment via the anonymization node
The pattern
Every one of these workflows shares the same core: before real data reaches an LLM, a node pushes it through anonymization – and pulls the result back to the original afterwards. That makes automation GDPR-compliant without giving up the quality of international frontier models.
What you should know about Workflows.
Frequently asked questions
Workflows is a visual no-code/low-code automation builder embedded directly into anymize. You drag building blocks (nodes) onto an infinite canvas and connect them with edges to form a flow. Each node has its own inspector panel for configuration. Workflows are executed by a Rust-based executor – fast, deterministic and robust.
We stand behind anymize. And we know – when an AI tool touches client, patient or employee data, a demo video isn't enough. That's why we give you 14 days of full access – all models, all features, no credit card. Enough time to be certain, before you trust us.
Your AI workplace awaits.