Privacy Policy

1. Who we are

Mad Science Software B.V. (trading as Mad Science Software), Joop Geesinkweg 601, 1114 AB Amsterdam-Duivendrecht, the Netherlands, Chamber of Commerce number 42084958, makes Chiron. For the personal data described in this policy we are the controller under the General Data Protection Regulation (GDPR). Contact us about privacy at inquiries@madsciencesoftware.dev.

2. Our role, and your team's

Memories your team keeps in Chiron, and the personal data in them (for example a client's name in an email), belong to your organization. For those, your organization is the controller and we are its processor under our Data Processing Agreement. If you want to know what is held about you in a team's memories, ask that organization; we will help it answer.

This policy covers what we decide on ourselves: accounts, billing, the website and the running of the service.

3. What stays on your computer

The Chiron app keeps every memory in a database file on your own computer (in the .chiron folder in your home folder). Turning text into searchable form, and searching it when an assistant asks a question, happen on your computer too. Without a team, we receive none of your memories and none of your questions.

The app does contact the internet for a few things, which reveal your Internet Protocol (IP) address to the site it contacts:

• once a day, it checks for a newer version, and on request downloads it, from GitHub;
• on first run, it downloads the program that computes embeddings from GitHub, and the embedding model from Ollama's model registry.

4. What reaches us when you use a team

When you connect a computer to a team, the app syncs with our team server:

Account details

Your email address, your name if your sign-in provides one, the date you joined, your role in each team, and invitations sent to or by you (the invited email address and role).

Team details

Your team's name, plan, seats and storage use, and the sharing policy your owner or admins write.

Devices

Each connected computer's name, its operating system, when it last synced, and a secret it signs in with, stored only in scrambled (hashed) form.

Private memories

Only if you turn on private backup in the app: the full text of each memory you keep private (for example an imported email or document), the short sentences (cues) that point to it with their authors and dates, the numeric fingerprints (embeddings) used to search them, and where it came from. Visible only to you, on your connected computers and in the dashboard. Private backup is off by default; then your private memories stay on each of your computers and we don't receive them. Private memories that an Enterprise connector makes, for example from Gmail, are made on the team server and stay there.

Shared memories

The same, for memories you or teammates choose to share. Visible to everyone on the team.

Activity records

When a memory was shared, unshared or deleted, and by whom; not its content. Changes to the team's sharing policy and plan, with who made them.

When your assistant runs on your computer, we do not receive the questions it asks or which memories it recalls.

If you choose to connect a web-based assistant such as claude.ai or ChatGPT, that assistant sends your questions to our servers, where we search the memories you may see (your team's shared memories and your own private ones that are on the team server) and send the matches back to the assistant. This runs on our servers in the United States, alongside the team server. We use each question only to answer it and don't store it. We keep a record of each connection (which assistant, who connected it, when it was last used) and the security tokens that let it in, until you disconnect it or leave the team. The assistant's provider handles your questions and the memories it receives under its own privacy terms.

5. Billing

When a team subscribes, our payment provider Stripe receives the team's name and the owner's email address, and collects the card, billing address and any tax identification number directly. We receive the subscription's status, seats and billing period, not card numbers.

6. The website

Signing in is handled by Google Firebase Authentication, which also sends the email that verifies your address and keeps you signed in using storage in your browser. The website has no advertising or analytics trackers. Our servers keep technical logs of failed requests (which page and what went wrong) to fix problems.

If you email us, for example through "Talk to us", we keep that conversation to answer it.

7. Why we use it (legal bases)

We use personal data:

• To provide Chiron to your team: running accounts, sync, the dashboard and billing (performing our contract with your organization, and our legitimate interest in serving its members).
• To keep Chiron secure and working, and to fix problems (legitimate interest).
• To keep invoices and tax records (legal obligation).
• To answer you when you contact us (legitimate interest, or steps before a contract).

We do not sell personal data, share it for advertising, or use your team's memories to train artificial intelligence (AI) models.

8. Who we share it with

We use these service providers, each only to run its part of Chiron:

Amazon Web Services (AWS)

Hosts the team server, its database and this website, in the United States (us-east-1 region).

Google Firebase Authentication

Sign-in and email verification.

Stripe

Payments, invoices, the billing portal and tax calculation (Stripe Tax).

GitHub

Distributes the app and its updates; sees the IP address of computers that check for or download them.

We may also disclose data when the law requires it, or to a buyer if our business is sold, under the same protections.

9. International transfers

Our providers store and process data in the United States. Where the GDPR applies, those transfers rely on the European Union–United States Data Privacy Framework (DPF) where the provider is certified under it, and otherwise the European Commission's Standard Contractual Clauses (SCCs).

10. How long we keep it

Memories you delete are erased from the team server when your computer next syncs; a deletion record (which memory, when and by whom, without its content) remains so other computers can erase their copies too. Database backups keep deleted data for 7 days.

When you turn off private backup, we delete the team server's copies of your private memories; each of your computers keeps its own.

Invoices and tax records are kept as long as tax law requires.

When a member is removed from a team or leaves it, their private memories for that team are deleted from the team server; what they shared stays with the team, shown as from a former member. When you delete your account, we delete your memberships, your private memories, your registered computers and your sign-in account. A team's data is deleted when the team is deleted, or 30 days after its subscription ends. Copies remain in database backups for up to 7 days. Copies already on people's computers stay there until they remove them: we can't reach a computer once it's disconnected.

11. Your rights

Depending on where you live, including under the GDPR and the California Consumer Privacy Act (CCPA), you may ask to see, correct, delete or receive a copy of your personal data, object to or restrict how we use it, and withdraw consent where we rely on it. You can already read your memories in the dashboard, and delete any memory you saved from the app (unsharing it first if it is shared).

For anything else, write to inquiries@madsciencesoftware.dev. We will confirm who you are, then answer within the time the law sets. If your request is about a team's memories, we will pass it to that organization and help it respond. You can delete your account yourself from the Account page in the dashboard, and download a copy of your memories there first.

You may also complain to a data protection authority; ours is the Dutch Data Protection Authority (Autoriteit Persoonsgegevens).

12. Security

Connections to our servers are encrypted (Hypertext Transfer Protocol Secure, HTTPS). Each team's data is kept apart inside the database itself, so a request on behalf of one team can't read another's. Device and invitation secrets are stored only in hashed form. The team server's database and its backups are encrypted at rest.

Memories on your computer are as safe as your computer: use its disk encryption and a strong sign-in.

13. Changes and contact

If we change this policy, we will update the date at the top, and tell account holders beforehand about changes that matter. Questions: inquiries@madsciencesoftware.dev.