Privacy Policy

This policy explains how Cartographer ApS ("we", "us") handles personal data when you use Buzzer: the web app at https://go.usebuzzer.com, the chat widget and email channels our customers put in front of their users, the help centers we host, the Buzzer Pocket mobile app, the API and MCP server, and this website.

It's written to be read, not just published. Where the legal framing needs unpacking, the margin notes do that.¹

Who's who

Three kinds of people touch Buzzer, and the rules differ for each.

Workspace users are the people on a team that signs up for Buzzer — agents, managers, admins. For your account data (name, email, login, settings, audit trail) we are the data controller.²

Customers' end users are the people who talk to a Buzzer workspace: they open the widget on a customer's site, email a support address routed into Buzzer, or browse a help center. For everything they share, the workspace that runs Buzzer is the controller and we are their processor — we handle that data only on the workspace's instructions and under our data processing terms.

Visitors to usebuzzer.com are reading a marketing site. We keep that simple: see Cookies.

If you're an end user and want to know what a company does with your conversations, ask that company first. We'll help them answer, but we don't decide how your data is used — they do.

What we collect

Account and workspace data

• Name, email address and login credentials (password hash, or the identity from a sign-in provider you use); session tokens.
• Workspace configuration: team members and roles, inboxes, labels, saved replies, help-center settings, widget appearance, playbooks and skills you teach Buzz, connector and tool configuration.
• Credentials you give Buzzer to act on your behalf — API keys for the REST API, OAuth tokens for MCP connectors and other integrations. These are encrypted before they're written to the database.³
• Audit and activity logs: who did what in the workspace and when. Buzz's own actions are logged the same way.

Customer content

This is the data your workspace generates in the course of talking to customers:

• Conversations and messages across every channel — widget, email, in-app — including attachments.
• End-user identities and profiles: whatever a person or your integrations provide (name, email, company, custom attributes), plus notes and facts your team or Buzz records about them.
• Knowledge-base articles, help-center content and source files uploaded for Buzz to learn from, including search indexes and embeddings derived from them.
• Technical metadata that arrives with a message: IP address, browser type, locale, timestamps, and the page the widget was opened from where your embed passes it along.

We don't decide what goes into a conversation — your customers and your team do. Please don't route data into Buzzer that you're not entitled to process, and don't use it to collect special-category data (health, biometrics and the like) unless you have a lawful basis and have agreed a DPA with us that covers it.

Usage and diagnostics

Server logs, error reports and performance metrics that describe how the Service behaves. These are operational — we use them to keep Buzzer up and fix what breaks. We don't run third-party advertising or behavioural analytics scripts in the app or on this site.⁴

Talking to us

If you email us, book a demo or file a support request, we keep that correspondence so we can follow up.

How we use it

• To run the Service — authenticate you, store and deliver conversations, send and receive email on your behalf, render your help center, sync with the connectors you've enabled.
• To run Buzz — generate replies and drafts, search your knowledge base, summarise threads, suggest articles, and carry out the actions you've allowed. See Buzz and AI processing for exactly what's involved.
• To keep things safe — detect abuse, enforce rate limits, investigate incidents, and secure accounts.
• To bill and administer accounts — invoices, plan limits, usage.
• To talk to you — service notices, security alerts, replies to your questions, and occasional product updates you can opt out of.
• To improve Buzzer — aggregate, de-identified usage patterns. We don't read your conversations to do product research, and we don't use your content to train foundation models.

We never sell personal data, and we don't share it with advertisers.

Buzz and AI processing

Buzz is the agent inside Buzzer. When it works on a conversation, the relevant material is sent to a large language model to produce a reply, a draft, a summary or a plan: the messages in the thread, matching excerpts from your knowledge base, the playbooks and skills you've written, and facts recorded about the customer.

• Models come from Google (Gemini), OpenAI and Anthropic, reached through Cloudflare's AI Gateway. Embeddings for search run on Cloudflare Workers AI. Where Buzzer only needs to choose between fixed options, such as suggesting an icon, it uses TypeSafe's Jev classifier, which sees just the title and description of what's being labelled. Under the commercial API terms we use with each provider, prompts and outputs are not used to train their models.⁵
• Buzz is built to act only within the permissions you configure. Replying automatically, calling a tool, updating a record — each is something you turn on, and anything you mark as needing approval waits for a human.
• Outputs are probabilistic. Buzz can be confidently wrong. Your workspace is responsible for reviewing what it's allowed to send and for telling end users, where the law requires it, that they may be talking to an automated system.
• Buzz isn't designed for, and shouldn't be configured for, automated decisions with legal or similarly significant effects on a person without human review. Our Terms say what that means for your configuration.

Sub-processors

These are the companies that process personal data on our behalf to deliver Buzzer. We vet them, bind them to data processing terms, and keep this list current.

We don't share customer data with any processor beyond this list without updating it first.

Connectors, tools and destinations you configure

Buzzer can connect to other systems on your instruction — MCP servers, your own APIs, CRMs, ticketing and billing tools, webhooks, the Buzzer MCP server you plug into your own AI assistants. When you enable one, data flows to it because you configured it to. We process that data only to make the connection you configured. What the destination does with it afterwards is between you and them — we're not a party to that agreement and this policy doesn't extend to it. Review a connector's own terms before switching it on, and remove it when you stop using it.

Retention

• Account and workspace data is kept for as long as your workspace exists. When you close a workspace we delete its data within 30 days, except where we must keep records (for example invoices) for longer by law.
• Customer content is kept until your team deletes it or the workspace is closed. Deleting a conversation, article, file or person removes it from the live Service; copies may persist in backups for up to 30 days before they age out.⁶
• Email we send and receive on your behalf is kept as part of the conversation it belongs to.
• Operational logs are kept for a short period — typically 30 days, up to 90 for security-relevant events — then discarded.
• Model traffic isn't retained by us beyond the conversation it produced. AI Gateway logs are kept for up to 90 days for debugging and cost accounting, then discarded.

Security

Buzzer is a small system with few moving parts, and we'd rather tell you exactly what those parts are than call them enterprise-grade.

• Tenant isolation. Every record is designed to be scoped to a workspace, with that scope checked on every request.
• Where your data lives. The primary database is hosted in Frankfurt, inside the EEA. Compute runs on Cloudflare's network, and attachments and files are stored there too. The only other parties designed to see customer content are the model providers listed under Sub-processors, and only the material Buzz needs for that reply.
• Encryption. Data is encrypted in transit (TLS) and at rest by our infrastructure providers. Connector credentials and API keys get a second layer of encryption of our own before they're written.
• Model providers. Our contracts with Google, OpenAI, Anthropic and TypeSafe commit them not to train on prompts or outputs sent through their paid APIs. See Buzz and AI processing.
• Buzz's guardrails. Buzz is built to act only within the permissions you configure. Anything you mark as needing approval waits for a human, and every action Buzz takes is written to the same audit trail as your team's.
• No trackers. No advertising scripts, analytics pixels or session replay in the app or on this site.
• Access inside the company is limited to the people who need it to run or support the Service, and access to production systems is logged.
• Deletion is soft first and purged on a schedule, so a mistaken delete can be recovered briefly and a real one leaves live systems immediately and backups within about 30 days. See Retention.

Reporting a vulnerability. If you find a security problem, email hello@usebuzzer.com and we'll acknowledge it within a few business days. Please ask us before testing the Service — our Terms require written permission first — and give us a reasonable chance to fix what you find before publishing it.

No system is perfectly secure. If we learn of a breach affecting your data we'll tell you without undue delay, with what we know and what we're doing about it, so you can meet your own obligations.

Cookies and local storage

• The app uses a session cookie to keep you signed in and a small amount of local storage for preferences such as theme. That's it.
• The widget stores an identifier in the visitor's browser so a conversation can continue across page loads. It's scoped to the site the widget is on and isn't used to track people around the web.
• This website sets only the strictly necessary cookies our infrastructure provider (Cloudflare) uses for security and bot protection. Nothing for advertising or analytics.

We don't use advertising or cross-site tracking cookies anywhere, so there's no consent banner to click through — the cookies above are the strictly necessary kind that doesn't need one.

International transfers

We're based in Denmark, and our primary database is hosted in Frankfurt, inside the EEA. Cloudflare runs a global network, PlanetScale operates that database from the United States, and the model providers' APIs are served from the United States, so personal data may still be processed outside the EEA. Where it is, we rely on the European Commission's Standard Contractual Clauses and, where the recipient is certified, the EU-US Data Privacy Framework, plus the technical measures above.

Your rights

If you're in the EEA or the UK you can ask us to access, correct, delete, restrict or export your personal data, and you can object to processing we base on legitimate interests. Email hello@usebuzzer.com and we'll respond within a month — the law lets us take up to two months more for complex requests, and we'll tell you if we need to. We're established in Denmark and haven't appointed a UK representative; UK requests come to the same address. You can also complain to your supervisory authority — ours is the Danish Data Protection Agency (Datatilsynet).

If you're an end user of a company that uses Buzzer, the quickest route is to contact that company: they control your data and can act on your request directly. If they ask us for help, we will.

Children

Buzzer is a business tool and isn't directed at children. We don't knowingly collect personal data from anyone under 16 for our own purposes. Our customers are responsible for how they use the Service with their own audiences.

Changes to this policy

We'll update this page when our practices change and bump the effective date at the top. For material changes — a new category of data, a new purpose, a new sub-processor — we'll notify workspace admins by email ahead of time.

Contact

Cartographer ApS, registered in Denmark under company number DK46298578. Æbeløgade 23, 3TV, 2100 København Ø, Denmark.

Privacy questions and data requests: hello@usebuzzer.com. A signed data processing agreement is available on request.

Footnotes

1. The footnotes on this page are explanations, not part of the binding text. If a note and the body disagree, the body wins. ↩

2. "Controller" is GDPR language for whoever decides why and how data is used. "Processor" means acting on someone else's instructions. Buzzer is both: controller for your team's account, processor for the conversations you have with your customers. ↩

3. Connector secrets are encrypted at rest with AES-GCM, using a key derived from a deployment secret rather than stored alongside the data. Rotating that secret invalidates stored credentials, which then have to be re-entered. ↩

4. No Google Analytics, no Meta pixel, no session replay. The only third parties that see requests are the infrastructure providers listed under Sub-processors. ↩

5. "Not used to train" is a contractual commitment each model provider makes for paid API traffic, not an inference we're drawing. If we add or change providers we'll update this section and the sub-processor list first. ↩

6. Most deletes in Buzzer are "soft" first — the row is stamped as deleted and hidden immediately, then purged on a schedule. That's what lets us recover a mistaken delete for a short window and still honour the 30-day limit. ↩