Privacy Policy

Last updated: 3 August 2026

This policy explains what Fernary does with personal data. It is written to be read, not to be survived.

Controller: Jeremiah Lena, trading as Fernary, a sole trader established in Ireland at Dublin, Ireland. Contact for anything in this policy: legal@fernary.com.


1. The two hats we wear

This distinction runs through everything below, so it comes first.

For your own account, we are the controller. Your email, your name, your workflows, your billing details when there are any — we decide how those are handled, and this policy is our account of it.

For the data your workflows pull in, we are your processor. When a workflow reads your Shopify orders, your Gmail inbox, or your Stripe customers, it is handling other people's personal data — your customers'. You are the controller of that data. We process it only to run the workflow you built. You are responsible for having a lawful basis to process it and for telling those people what you do with it.

If you need a Data Processing Agreement to cover that relationship, email legal@fernary.com and we'll put one in place. We haven't published a standard one yet.

2. What we collect

Account data — because you signed up

DataWhy
Email addressIdentifies your account; receives sign-in codes and service notices
Name and avatar URLShown in the app. Only if you sign in with Google, which supplies them
Google account IDLinks your Google sign-in to your account
Sign-in codesStored hashed only, expire quickly, dead once used or after too many wrong attempts
Session tokensHeld in Redis so you stay signed in

We do not store a password, because there isn't one.

Content you create — because that's the product

Workflows, prompts, AI-builder conversations, chat sessions, run history, and the contents of your Data Stores.

Run history is the sensitive part. A run record keeps each step's output, and those outputs are whatever your workflow fetched — an order, an email body, a customer's name and address. If your workflows touch personal data, your run history contains personal data.

Credentials for connected accounts

Access and refresh tokens for the services you connect. These are encrypted at rest before they reach the database. Any LLM API keys you enter yourself stay in your browser's local storage and are never sent to our servers.

Technical data — because the software has to run

Server logs and operational telemetry: request paths, timings, error messages, workflow and run identifiers, IP address, and browser user-agent. Log output is filtered to strip values whose names look like secrets — tokens, passwords, keys, authorization headers, signatures.

What we don't collect

  • No analytics, no tracking pixels, no advertising, no third-party cookies. There is no PostHog, no Google Analytics, no Meta pixel, no session recorder.
  • No cookie banner, because the only browser storage we set is what signs you in and remembers your preferences — strictly necessary, which needs no consent.
  • We never buy personal data, and we never sell yours.

3. Why we're allowed to (lawful bases)

PurposeLawful basis
Providing the Service, running your workflowsContract — Art. 6(1)(b)
Sign-in, session security, abuse preventionContract and legitimate interests — Art. 6(1)(b), (f)
Keeping the Service working: logs, debugging, capacityLegitimate interests — Art. 6(1)(f)
Service emails you need to receiveContract — Art. 6(1)(b)
Billing and tax records, once there are feesContract and legal obligation — Art. 6(1)(b), (c)
Any marketing emailConsent — Art. 6(1)(a), withdrawable in one click

4. Who your data reaches

Everything here is a processor or sub-processor acting on our instruction, or a service you chose to connect. None of them get your data to use for their own purposes.

To run the Service

WhoWhat reaches themWhere
RailwayHosting, PostgreSQL, Redis — so, everything at restEU (Ireland) — eu-west-1
VercelServes the web app; sees request metadata and IPGlobal edge
ResendRecipient address and message content for emails we send, including sign-in codesUS/EU

AI providers — only what a workflow sends them

WhoWhen
Anthropic (Claude)An AI step or a branch condition using a Claude model
OpenAIAn AI step using a GPT model
Google (Gemini)An AI step using a Gemini model
xAI (Grok)An AI step using a Grok model
Brave SearchA web-search step — your search query
Jina ReaderA page-read step — the URL to fetch

These providers do not train on your content. Anthropic and OpenAI do not train on data submitted through their APIs by default, and we do not opt in. Prompts may be retained briefly for abuse monitoring under their own terms.

Only what a step actually needs is sent. A prompt containing customer data means that customer data goes to the model provider. That is inherent to what an AI step is, and it is worth thinking about when you write one.

Services you connect

Notion, Linear, GitHub, GitLab, Gmail, Stripe, Shopify, Google Calendar, Google Drive, Google Docs, Google Sheets, Outlook, Slack. We exchange data with these only as your workflows instruct. Each has its own privacy policy, which governs what they do at their end.

In your browser

Loading the app fetches integration logos from logo.dev and Google's product logo CDN (gstatic.com). Those requests reveal your IP address and user-agent to those hosts. No account data goes with them.

Otherwise

We disclose personal data only where legally compelled, to establish or defend legal claims, or to a successor if the business transfers — including on incorporation. We'd tell you about a transfer first.

5. International transfers

Your data at rest stays in the EU. The database, the cache, and the stored contents of your workflows and runs are hosted in Ireland (eu-west-1) and are not replicated outside the EEA.

Data does leave the EEA in two situations, both in transit:

  • When a workflow calls an AI provider. Anthropic, OpenAI, Google, xAI, Brave, and Jina process the request wherever they operate, typically the US. We store nothing there.
  • Serving the app and sending email. Vercel serves from a global edge network, and Resend may process outbound mail in the US or the EU.

For those, we rely on the European Commission's adequacy decision for providers certified under the EU–US Data Privacy Framework, or on Standard Contractual Clauses where they are not. Email legal@fernary.com and we'll tell you which applies to a given processor.

If keeping a particular workload inside the EEA matters to you, say so — some model providers offer EU-only processing, and we can talk about which steps to route where.

6. How long we keep it

DataRetention
Account dataWhile your account exists
Workflows, Data Stores, chatsWhile your account exists, or until you delete them
Run history and logsWhile your account exists, or until you delete the run or workflow
Sign-in codesMinutes. Deleted once used, expired, or exhausted
SessionsUntil expiry or sign-out
Operational telemetryShort rolling window, then aged out
Billing records, once there are anyAs long as tax law requires — typically 6 years

When you delete your account, we delete your account data, workflows, run history, Data Stores, chat history, and connected-account tokens within 30 days, apart from anything we must keep by law. Backups age out on their own cycle, so a copy can persist a little longer before it is overwritten.

Deleting a workflow deletes its runs. Disconnecting an integration deletes its stored token.

7. Your rights

Under the GDPR you can ask us to:

  • Access — get a copy of your personal data
  • Rectify — correct it
  • Erase — delete it
  • Restrict or object to processing, including anything based on legitimate interests
  • Port — receive it in a machine-readable form
  • Withdraw consent — where consent was the basis, without affecting what came before

Email legal@fernary.com. We'll respond within one month. There's no charge unless a request is excessive, and we'll say so first if it is.

You can also complain to the Irish Data Protection Commission (dataprotection.ie) or to your local supervisory authority. We'd rather you told us first, but it's your right either way.

If you're one of our customers' customers — your data reached us because a Fernary user's workflow processed it — we are the processor, not the controller. Send your request to that business. If you reach us instead, we'll forward it and tell you we have.

8. Automated decision-making

Fernary is a tool for building automations, so your workflows may make automated decisions. We don't make automated decisions about you.

If a workflow you build makes decisions about people with legal or similarly significant effects, Art. 22 GDPR obligations are yours as controller. The approval step exists partly for this: it puts a human in the loop where you need one.

9. Security

  • OAuth tokens encrypted at rest; sign-in codes stored only as hashes
  • HTTPS everywhere in transit
  • Passwordless sign-in, so there is no password to breach or reuse
  • Every query scoped to the authenticated user; tenant isolation tested
  • Secret-looking values stripped from logs by name
  • Outbound requests blocked from reaching private network addresses

No system is perfectly secure, and this one is in beta. If a breach affects your personal data we'll notify the DPC within 72 hours where required, and you without undue delay where the risk to you is high.

Found a vulnerability? security@fernary.com. We won't pursue good-faith research.

10. Children

Not for under-16s. We don't knowingly collect their data; tell us at legal@fernary.com if you believe we have and we'll delete it.

11. Changes

We'll update this page and the date above. For changes that materially affect your rights we'll email you at least 14 days before they take effect.


Data protection questions: legal@fernary.com. Security: security@fernary.com.