Privacy
Privacy policy
What personal data Rykon AI collects, why, the legal basis, who else handles it, how long it is kept and the rights you have.
- Version
- 1.5
- Effective
- 8 November 2026
- Last updated
- 8 November 2026
1.Who we are and our two roles
1.1Rykon AI is provided by Rykon Technologies Inc. ("we", "us"). We handle personal data in two different roles, and this policy explains both.
- As Controller
- For Account Data (the people who sign up for and use Rykon AI), and for visitors to our websites and people who contact us. This policy is our notice to you for that data.
- As Processor
- For Customer Data: the tasks our customers give their agents, the material in them, and what their runs produce. Our customer decides how that data is used and is its Controller; we process it only on the customer’s instructions under our data processing addendum.
2.What we collect as Controller
Data
Name, work email and organisation
- Where it comes from
- You, at sign-up
- Why we use it
- To create and run your account
- Legal basis (GDPR)
- Contract
Data
Sign-in details: password (held only by our identity service), passkeys or authenticator apps you add, email confirmation
- Where it comes from
- You
- Why we use it
- To keep your account secure
- Legal basis (GDPR)
- Contract; legitimate interests in security
Data
If you choose Sign in with Google: your Google account ID, email address, whether Google has verified it, and your name (see Sign in with Google)
- Where it comes from
- Google, when you ask it to sign you in
- Why we use it
- To create your account and sign you in
- Legal basis (GDPR)
- Contract
Data
Your role, and what you and your team do in the console and with keys: who changed what, when, and every refused request (the audit log)
- Where it comes from
- Your use of the service
- Why we use it
- Security, accountability and support
- Legal basis (GDPR)
- Legitimate interests; legal obligation where records are required
Data
Usage: runs, model calls, tokens and their cost
- Where it comes from
- Your use of the service
- Why we use it
- To show you usage, apply spend ceilings and, once plans open, to bill
- Legal basis (GDPR)
- Contract
Data
Support requests: your message and your account
- Where it comes from
- You
- Why we use it
- To answer you
- Legal basis (GDPR)
- Contract; legitimate interests
Data
Technical data such as IP address and browser, in server logs
- Where it comes from
- Your browser
- Why we use it
- To run, secure and debug the service
- Legal basis (GDPR)
- Legitimate interests in security
2.1Our identity service at id.rykon.ai is run by us and is the same one that signs people in to our other products, such as Rykon Mail. If you already have an account on one of them under the same address, you are asked to sign in rather than sign up again. What you may do in Rykon AI is decided by Rykon AI’s own records, never by another product.
2.2Where we rely on legitimate interests, we have weighed them against your rights; you may object at any time (section 9).
3.What we process for our customers
3.1As Processor, and only to provide the service, we process:
- tasks and agent briefs, and any personal data a customer puts in them;
- transcripts of runs: what the model said, the tools it called, their inputs and results, and the run’s result. Anything shaped like a key or token is redacted before a transcript is stored;
- findings and other output a run records;
- approvals: the action an agent proposed, any edit a person made, who decided and when;
- the record: every model call (the model, tokens and cost, never the content), tool call, approval step, agent change and kill switch.
4.AI models
4.1Agents run on AI models provided by Anthropic, reached through an AI gateway and model router we run ourselves; if Anthropic does not answer, the router may send the call to OpenAI instead. A model call carries what the run needs: the task, the agent’s brief, the conversation so far and tool results. It never carries your account’s gateway key, your API keys or a run’s token: the sandbox does not hold them, and our relay adds the gateway key on its way out.
4.2Our gateway keeps a usage record of each call (account, model, tokens and cost), and our model router keeps a trace of calls, including their content, in a tracing tool we host ourselves, to diagnose faults. We do not use Customer Data to train AI models, and we use the providers’ commercial API services, not their consumer products.
6.International transfers
6.1Rykon AI’s servers, databases and sandboxes are currently hosted in Germany; as we open more regions, each is listed on our subprocessors page before it holds customer data. The AI model providers, Anthropic and OpenAI, are in the United States, and account emails may be delivered through Amazon SES in the United States. Where personal data from the European Economic Area, the United Kingdom or Switzerland is transferred to a country without an adequacy decision, the transfer relies on the European Commission’s Standard Contractual Clauses (with the UK International Data Transfer Addendum for UK data) in the provider’s data processing terms, or on the provider’s certification under the EU-US Data Privacy Framework where it holds one. You may ask us for a copy of the relevant safeguards.
7.How long we keep it
Data
Account Data
- Kept for
- While the account is open. Erased 7 days after an owner asks to close the account
Data
Runs, transcripts, findings, approvals, tasks, knowledge, connections and mail inboxes
- Kept for
- While the account is open. Deleted 7 days after an owner asks to close the account
Data
A person removed from an account
- Kept for
- Their name, email address and sign-in identity are erased 30 days after removal, everywhere the account’s data holds them, or at once on a privacy request
Data
The record and the audit log
- Kept for
- While the account is open. Both are append only and hash chained: nobody can edit or delete an entry. When a person or an account is erased, the entries that named them keep only what happened and when, with names, addresses and content replaced, and the chain still proves that nothing else changed
Data
Mission Control conversations (what you and Rykon said, as text)
- Kept for
- 90 days, unless you turn keeping them off (then only counts and timings are kept); you can delete one at any time. Audio is never stored by us
Data
Background job records
- Kept for
- 90 days after the job finished
Data
Sign-in session
- Kept for
- 90 days, or until you sign out; a sign-in attempt, 10 minutes. Sign-in requests and codes for apps connected over MCP, 1 day after they expire; their access tokens, 30 days after they expire or are revoked
Data
Encrypted backups of our databases
- Kept for
- Rotated on a rolling schedule; erased data leaves them within four months
7.1These periods are enforced by the service itself, every day. Owners and admins can download everything an account holds at any time (Settings, Your data), and before it closes.
8.Security
8.1We protect personal data with technical and organisational measures appropriate to the risk, described on our security page: a fresh, isolated sandbox for every piece of work that runs code, a shell or files, a relay that is the only address a sandbox can reach, keys held only as hashes, the gateway key sealed and never given to a sandbox, accounts walled off from one another, approvals by a person, and append-only records. No system is perfectly secure; we tell you of a breach as section 11 describes.
9.Your rights
9.1Depending on where you live, you have the right to access the personal data we hold about you; to correct it; to have it deleted; to restrict or object to its processing, including processing based on legitimate interests; to receive it in a portable format; and to withdraw consent where we rely on it. California residents have the rights to know, delete, correct and opt out of sale or sharing (we do neither), and not to be discriminated against for using them.
9.2To exercise a right, write to support@rykon.ai with "Privacy request" in the subject. We acknowledge a request within 10 days and answer within one month (45 days for California requests), which we may extend where the law allows and will tell you if we do. We may ask you to confirm your identity first. An authorised agent may act for you with your signed permission.
9.3If your data is in a customer’s tasks or runs, that customer is the Controller: contact them first. See your privacy rights for how, and what we do to help.
9.4You also have the right to complain to a data protection authority, in particular in the EU member state where you live or work, or to the UK Information Commissioner’s Office. We would welcome the chance to address your concern first.
10.Children
10.1Rykon AI is a service for businesses and is not directed to children. We do not knowingly collect personal data from anyone under 16. If you believe a child has given us personal data, contact us and we will delete it.
11.Personal data breaches
11.1If a breach affects personal data we control, we notify the competent supervisory authority without undue delay and, where feasible, within 72 hours of becoming aware of it, and the people affected where the breach is likely to result in a high risk to them. For Customer Data, we notify the customer as the data processing addendum sets out.
11.2To report a suspected breach or a security problem, write to support@rykon.ai with "Security" in the subject. Breach notices from us come from that address too.
13.Sign in with Google
13.1You can create your Rykon AI account, or sign in to it, with your Google account instead of a password. It is your choice, and an account made with a password can always be used without Google.
13.2When you choose it, Google asks for your permission and then tells our identity service (the Rykon account, run by us) your Google account ID, your email address, whether Google has verified that address, and your name. We ask Google for nothing more: only the basic “email” and “profile” permissions. We cannot read your Gmail, contacts, files or calendar, and we never post or act on your Google account. Google never sends us your Google password.
13.3We keep your Google account ID with your account so we recognise you the next time you sign in, and your name and email become your account’s name and email. Google also gives our identity service a sign-in token that expires after about an hour; it is kept only with that link to your account and is not used for anything else. We use what Google shares only to sign you in and run your account, never for advertising or to train AI models, and we do not sell it or share it with anyone.
13.4An address Google has verified counts as confirmed, so an account made with Google needs no confirmation email. If you already have an account with the same address, it is joined to your Google account only when you have confirmed that address with us; otherwise sign in with your password instead. Staff accounts never sign in with Google.
13.5You can stop using Google at any time: set a password with “Forgot username or password?” on the sign-in page, and remove Rykon AI from your Google Account under Security, third-party connections. The link is deleted with your account.
13.6Rykon AI’s use and transfer of information received from Google APIs adheres to the Google API Services User Data Policy, including the Limited Use requirements.
14.Contact and changes
14.1Privacy questions and requests: support@rykon.ai, subject "Privacy request". We will tell you by email, and on this page, before a material change to this policy takes effect.
Change log
Version 1.5
8 November 2026Mission Control: what you and Rykon say is kept as text for 90 days unless you turn keeping it off; the voice is provided by OpenAI and is never recorded by us.
Version 1.4
8 November 2026Sign in with Google: what Google shares with us when you choose it, what we keep and why, how accounts are joined, how to stop using it, and the Limited Use commitment.
Version 1.3
5 November 2026Published how long each kind of data is kept, how a person and a closed account are erased, and where to report a breach.
Version 1.2
1 November 2026Said that the sandbox is for work that runs code, a shell or files; simple answers come from our control plane.
Version 1.1
31 October 2026Said where data is hosted today, and that a new region is listed on the subprocessors page before it holds personal data.
Version 1.0
30 October 2026First published.