Privacy Policy

Version 3.0 · Effective 6 October 2026

1. Who We Are

CashCurve AI is operated by Cash Curve. Our contact email is support@cashcurve.ai.

2. What We Collect

We collect the following categories of data:

• Account information: email address, name (provided at signup or waitlist)
• Uploaded content: financial statements (CSV, Excel, PDF) you choose to upload
• Structured financial data: income, expenses, assets, and other profile data you input
• AI assistant messages: conversations with the CashCurve AI assistant
• Operational data: content-free feature totals, necessary technical diagnostics, a separate private AI cost ledger, and structured product-feedback records derived from chat. Feedback jobs, checkpoints, session identifiers, message fingerprints and reporter correlation are retained operational records, not anonymous data.
• Location declaration: your self-declared country of residence (Australia or Outside Australia), recorded at sign-up

3. How We Use Your Data

• Provide the Service: process your data, generate projections, power the AI assistant
• Maintain and improve: fix bugs, improve parsing accuracy, and identify commonly requested features using content-free feature totals, bounded technical diagnostics and configured background extraction of product feedback from recent chat (see Section 4)
• Security: detect and prevent fraud, abuse, or security threats
• Legal compliance: meet our obligations under Australian law
• Geographic disclosure: display appropriate jurisdiction notices based on your declared country of residence, and maintain an audit record of your acknowledgment that the Service uses Australian financial rules

4. How We Improve the Service

Local feature totals count selected ordinary operations without document, chat, financial or actor content. They are partial operational counts, not complete journeys. Necessary error diagnostics are separately limited to approved technical fields. Optional PostHog analytics runs only after you choose it. It measures page visits, feature use and completion or failure using random identifiers. Browser IP is used for approximate country and region, then discarded. Financial values, documents, chat contents, names and email addresses are not sent. Automatic capture and session replay are disabled. The private AI cost ledger is separate: actor and provider references can identify an account and are restricted operational records, not anonymous analytics.

The service also supports background extraction of product feedback from recent saved chat messages to identify bugs, feature requests and usability problems. When enabled and executed, this sends a recent conversation window to OpenAI; it can contain financial information or discussion of documents. This first-party product-improvement purpose is separate from content-free measurement, necessary diagnostics and provider model training. Source configuration defaults feedback extraction to enabled, but deployed settings and actual worker execution require verification.

We retain structured feedback titles and descriptions, processing jobs and checkpoints, and session, message-fingerprint and reporter-correlation records. These records can remain identifying even when direct names are not stored. Authorised access, effective retention and deletion, historical handling and provider account controls require separate release evidence; we do not claim a fixed retention period, automatic unlinking or completed historical purge.

5. AI Processing Disclosure

AI processing can send document content, extracted financial information and chat messages to OpenAI. Response storage settings, files, vector stores, containers, caches, background product-feedback extraction and Agents tracing have distinct purposes, handling and retention controls. A request setting such as store=false does not establish universal zero retention, no training or Australian-only processing. Provider account settings, locations and historical copies require separate verification; we make no blanket guarantee for those channels.

Only upload your own household material with every affected adult aware of and authorising processing. One login can hold multiple profiles, and anyone using that login can access all household data; separate partner permissions and revocation are not available. Manually remove unnecessary TFNs, signatures, employer/client details and unrelated-person material before upload, keeping necessary balances, dates and transactions. There is no automatic redaction. The founder checks document families, both-adult authority and minimisation before real onboarding. Unsuitable material is deferred or given an agreed manual path; do not send real qualification files by email or chat.

If you use voice dictation in chat or the feedback form, speech recognition is handled by your browser or operating system speech service. CashCurve receives the resulting text only when you choose to send it, and we do not receive or store your microphone audio.

6. Third-Party Processors

The following channels can process data. Locations, training use and retention depend on deployed configuration and provider terms; the production evidence ledger must verify them before cohort release. Unknown values are not guarantees.

6A. Geographic Scope & Jurisdiction

The Service is designed for Australian residents and uses Australian tax rules, superannuation systems, property data sources, and financial assumptions. During sign-up, we ask you to declare whether you are located in Australia or elsewhere. We record:

• Your declared country of residence (Australia or Other)
• The date and time of your declaration
• For users outside Australia: the date and time you acknowledged that the Service uses Australian financial rules

This information is used to:

• Display appropriate notices about the Service's geographic scope
• Maintain an auditable record of informed consent
• Comply with consumer protection obligations under Australian law

Your declared country is stored as a simple category (Australia / Other) and is not considered sensitive personal information. It is not encrypted at the application layer but is protected by database-level encryption and access controls.

If you are located outside Australia, please be aware that all calculations, projections, and AI-generated insights are based on Australian financial systems and may not be accurate for your jurisdiction. Your data is processed and stored in accordance with Australian law. You may have additional rights under your local privacy laws (such as the EU General Data Protection Regulation or UK Data Protection Act); contact us at support@cashcurve.ai to exercise these rights.

7. Cross-Border Disclosure

The primary application deployment targets AWS Sydney. Authentication, AI processing, delivery, monitoring, speech and recovery infrastructure can involve processing outside Australia. This is not an Australian-only data residency guarantee; deployed locations require release verification.

If you are located outside Australia, your use of the Service constitutes a voluntary cross-border transfer of your data to Australia. We process your data under Australian privacy law. If your local jurisdiction provides additional data protection rights (e.g., GDPR right to erasure, data portability), contact us and we will honour reasonable requests.

8. Cookies & Tracking

We use cookies or similar storage for necessary functionality and preferences. Optional PostHog analytics requires your separate opt-in. You can withdraw it at any time through Cookie Preferences. Local content-free totals, necessary diagnostics and configured chat-derived product feedback are separate purposes; withdrawing consent stops new measurement but does not delete historical provider copies. For full details, see our Cookie Policy.

9. Data Retention

• Uploaded originals: removed from our active file storage within seven days of upload and may be removed sooner. You can upload a document again if needed.
• Extracted facts: remain in your account after originals are removed. Reset clears active financial data and keeps your login and structured Settings backups. Backups do not include original uploaded files.
• Deliberate product feedback: text and optional reviewed page images and technical details are kept in active support storage for up to 30 days from the first send, or removed earlier when you delete your account. Access is restricted to permissioned support administrators. They are excluded from financial backups and AI processing. Infrastructure recovery backups follow the separate schedules below.
• Diagnostics and provider copies: separate schedules require production and historical-buffer verification; upload expiry does not establish their deletion.
• Product feedback: retained chat-derived titles/descriptions, jobs/checkpoints and session/message/reporter correlation have separately unverified access, retention, deletion and historical handling. Account deletion is not proof of erasure or unlinking of every derived record.
• Operational records: content-free aggregate totals are distinct from restricted actor-linked AI spend and product-feedback records. The planned allowance ledger retention window is the current and prior two completed UTC months, subject to implementation and release verification. Unsettled usage may require a detached minimum liability record after deletion; this is not a guarantee that every linked record is erased immediately.

Deleting your account removes sign-in access and active account data, including Settings backups. File cleanup may still be finishing after you are signed out. Linked processing files are also queued for deletion.

Infrastructure recovery backups and provider recovery copies follow separate retention schedules and are not erased immediately by account deletion. Contact support@cashcurve.ai for the applicable schedules and any retained manual or final backup exceptions. We do not promise deletion of provider recovery copies beyond the provider controls that apply to them.

10. Security Measures

We implement multiple layers of security to protect your data:

• Document encryption: uploaded files are encrypted with AES-256-GCM before they are stored
• Database encryption: selected direct identifiers (names, emails, account numbers, addresses) is encrypted at the application layer before being written to the database
• Automatic document deletion: uploaded document originals are removed from active file storage within seven days of upload and may be removed sooner
• Staff access: authorised staff can access financial descriptions and operational records. Encryption does not make amounts, descriptions or actor-linked rows non-identifying.
• Transport protection: HTTPS protects the application connection; effective provider and infrastructure transport settings require release verification.
• Monitoring & incident response: bounded technical diagnostics and documented incident procedures; this is not a promise of continuous observation of every failure.

11. Your Rights

You have the right to:

• Access your personal data
• Correct inaccurate data
• Request deletion of your account, subject to the cleanup, recovery and operational-record limits above
• Export your data
• Complain to the Office of the Australian Information Commissioner (OAIC)

12. Children's Privacy

The Service is not directed at individuals under 18 years of age. We do not knowingly collect personal information from children.

13. Changes to This Policy

We may update this Privacy Policy from time to time. We will notify you of material changes via email or in-app notification. Financial access requires explicit acceptance of the current Terms and Privacy versions; browsing or continued use does not substitute for acceptance.

14. Contact & Complaints

For privacy enquiries, contact us at support@cashcurve.ai.

If you are unsatisfied with our response, you may lodge a complaint with the Office of the Australian Information Commissioner (OAIC).