Privacy Policy
How Duckk, Inc. collects, uses, shares and protects your data across duckk chat, the API gateway and this website — in plain language first, legal detail second.
At a glance
Duckk, Inc. (“Duckk”, operating duckk.org) offers a consumer AI assistant and a developer LLM API gateway, made for Africa. We collect only what we need to run your account, conversations, billing and security. We never sell your personal data, we do not use your content to train models, and every third party that processes data for us is listed below.
Contents 13 sections
1. Controller & scope
In short: Duckk, Inc. (privacy@duckk.org) is the controller of your account and usage data. For data you send through the API on behalf of your own end-users, you are the controller.
This policy covers duckk.org, the duckk chat assistant (web and any apps), guest access, the API gateway at api.duckk.org/v1, and related emails and support. The controller is Duckk, Inc. — reach us at privacy@duckk.org.
Developer note: if you send your own end-users' data through the API, you are the controller of that data and Duckk acts as your processor. You must give those users a privacy notice that covers processing by Duckk and the selected upstream model provider.
This policy does not cover the consumer products of upstream providers (for example the ChatGPT or Claude apps), which have their own privacy policies.
2. Data we collect
In short: Account profile, conversations and files, usage and billing, attribution and technical data — nothing more.
| Category | Examples | Source |
|---|---|---|
| Account & profile | Name, email, avatar, sign-in method (password, OTP, Google/Facebook/Microsoft), MFA status | You, via Clerk |
| Conversations | Chat messages, prompts, AI responses, feedback ratings, agent-mode steps, attachment metadata | Your use of duckk |
| Files | Original uploads in Backblaze B2; extracted text and chunks in MongoDB for full-text search | Your uploads |
| Usage & quotas | Token counts, estimated cost, message/attachment/storage quota usage | Automatic |
| Billing | Plan, payment history, transaction references from PayPal/Flutterwave/M-Pesa/MTN MoMo/Orange Money | You + payment processors |
| Attribution & referrals | UTM/gclid/fbclid, referrer, referral code, session ID (localStorage key duckk.attribution.v1) | Your browser |
| Analytics (optional) | Firebase Analytics (GA4) events; the user ID is SHA-256 hashed before reporting | Your device, if enabled |
| Technical context | Timezone, locale, coarse country/city from our CDN, device and browser type | Automatic |
| Security | Turnstile verification result, rate-limit and audit logs, salted SHA-256 hash of IP in guest mode | Automatic |
Note: We do not collect government IDs, precise geolocation or special-category data. In guest mode we store only a salted SHA-256 hash of your IP — never the raw IP.
3. How we collect it
In short: Directly from you, automatically from your browser and our infrastructure, and from providers like Clerk and payment processors.
Directly from you: sign-up, conversations, uploads, checkout, feedback and support emails.
Automatically: server and CDN logs, quota metering, the first-visit attribution snapshot stored in your browser's localStorage (duckk.attribution.v1 — UTM parameters, gclid, fbclid, referral code, entry path), and optional Firebase Analytics.
From our providers: Clerk syncs your profile when you sign in; payment processors confirm transactions; Cloudflare Turnstile returns a human-verification result.
Note: Voice input and read-aloud run in your browser via the Web Speech API — your audio is not uploaded to duckk servers.
4. Why we use it & legal bases
In short: Every purpose maps to a GDPR lawful basis: contract, legitimate interest, consent or legal obligation.
We do not use your conversations, prompts or files to train any model — see section 5.
| Purpose | Data used | Legal basis |
|---|---|---|
| Provide the assistant & API | Account, conversations, files, usage | Contract |
| Billing & quota enforcement | Plan, payment history, usage meters | Contract + legal obligation |
| Security & abuse prevention | Logs, Turnstile result, rate limits, IP hash | Legitimate interest |
| Product improvement | Aggregated usage, feedback ratings | Legitimate interest |
| Attribution & referral rewards | UTM/referral data, session ID | Legitimate interest |
| Analytics | Hashed analytics events | Consent (optional toggle) |
| Service emails | Email, locale | Contract |
| Legal & tax compliance | Billing records, audit logs | Legal obligation |
5. AI processing & model providers
In short: Your messages go through our gateway to the model you picked. No one trains on your data. Search, parsing and sandbox helpers see only what they need.
When you chat, your message and conversation context are routed through our self-hosted gateway to the upstream provider of the model you selected (OpenAI, Anthropic, Google, Meta, Mistral, DeepSeek, Moonshot/Kimi and others in the catalogue). The provider processes this content to generate a response.
No training on your data: we do not use your content to train any model, and upstream providers are instructed — by contract or API configuration — not to train on traffic routed through duckk.
Features that call specialist services: file parsing sends your document to MinerU for text extraction; web search sends your query to Brave or Exa; code execution runs in an isolated E2B sandbox; image and presentation generation go to the relevant image model. Each receives only the data needed for that task.
Speech features use your browser's Web Speech API; audio is not sent to duckk.
8. International transfers
In short: Made for Africa and served from African points of presence, but model inference and some subprocessors are in the US/EU — covered by SCCs.
duckk is made for Africa and served through African points of presence, but some processing happens abroad: model inference typically runs in the provider's US or EU region, and several subprocessors listed above are US-based.
Where the GDPR, POPIA, NDPA or the Kenya Data Protection Act applies, international transfers are safeguarded by Standard Contractual Clauses (SCCs) or equivalent mechanisms with our subprocessors. You can request a copy of the relevant safeguards at privacy@duckk.org.
9. How long we keep data
In short: You control your conversations; deleting them removes them permanently after 30 days. Billing records are kept as the law requires.
When you ask us to delete your account, we delete your account data — conversations, files and profile — and retain only the billing and compliance records the law requires.
| Data | Retention |
|---|---|
| Conversations & files | Until you delete them; then soft-deleted for 30 days and permanently deleted afterwards (including files in Backblaze B2) |
| Account profile | Until you delete your account; account data (conversations, files, profile) is then deleted |
| Usage records (tokens, cost) | 400 days |
| Monthly usage aggregates | 7 years (tax & financial compliance) |
| Audit logs | 2 years |
| Failed request logs | 30 days |
| Payment & billing records | As required by law (typically 7 years) |
| Guest-mode data | Short TTL; only a salted IP hash is stored |
| Attribution snapshot | Until you clear browser storage |
10. How we protect data
In short: TLS in transit, AES-256-GCM for sensitive fields at rest, rate limiting, Turnstile and audit logs.
We use TLS in transit, AES-256-GCM encryption for sensitive fields at rest, rate limiting, Cloudflare Turnstile bot protection, least-privilege access and audit logging of administrative actions.
No method of transmission or storage is 100% secure. If a breach is likely to affect your rights, we will notify you and the competent authority as the law requires. Report vulnerabilities to privacy@duckk.org — we acknowledge reports within 48 hours.
11. Your rights
In short: Access, correct, export or delete your data — much of it directly in the product. Email privacy@duckk.org and we answer within 30 days.
Under laws such as the GDPR, South Africa's POPIA, Nigeria's NDPA and the Kenya Data Protection Act, you may have the right to access, rectify, erase, restrict or object to the processing of your personal data, to data portability, to withdraw consent, and to lodge a complaint with your data-protection authority.
Built into the product: you can export any conversation and delete conversations or your whole account in settings. Deleting your account deletes your account data (conversations, files, profile); billing records are retained as required by law.
For anything else, email privacy@duckk.org from your account email. We respond within 30 days. Reasonable requests are free; manifestly unfounded or excessive requests may be declined with an explanation.
12. Children's privacy
In short: duckk is for users 16 and older; we do not knowingly collect data from children.
duckk is not directed at children. You must be at least 16 (or the age of digital consent in your country, if higher) to use the service. We do not knowingly collect personal data from children; if you believe a child has provided data, contact privacy@duckk.org and we will delete it promptly.
13. Changes to this policy
In short: Material changes are announced by email and site banner at least 14 days before they take effect.
We may update this policy as the service evolves. For material changes we will notify account holders by email and show a banner on duckk.org at least 14 days before the new version takes effect. Prior versions are available on request.
Privacy questions or requests?
For access, deletion, DPA/SCC requests or any privacy concern, email our privacy contact. For legal or terms questions, write to legal@duckk.org.
privacy@duckk.orgThis policy is provided for transparency and does not constitute legal advice. The English version at duckk.org/privacy is the authoritative version; translations are provided for convenience only.