On this page 12 sections
At a glance
- Accounts and usage connect to the Ocacity API; configured payments use Paddle. AI generation and remote deployment remain unconnected.
- Projects and bounded conversation history are saved to your account. Landing prompts can also appear in the URL.
- Account forms send credentials to the API. Sessions use HttpOnly cookies; passwords and session tokens are not stored in browser local storage.
This overview helps you navigate. The full document follows.
01Who is responsible for your information
AURTIZ LTD operates Ocacity and is the organisation responsible for deciding why and how personal information is used in connection with the service described here. This notice covers the Ocacity landing page, demonstration workspace, and preview authentication screens.
- Operator
- AURTIZ LTD
- Company number
- 17307727
- Registered office
- Office 19819, 182–184 High Street North, East Ham, London, E6 2JA, United Kingdom
- Legal & privacy contact
- Write to our registered office, marked Legal or Data Protection
This is a review draft describing the present preview. This notice is scoped to the preview described below. Production hosting, logging, and any new providers will require an updated notice before public launch.
This notice does not cover a separate website that someone builds from a Ocacity export. Its operator must provide their own privacy information.
02Information used by the preview
Information comes from what you type or select, records created by the workspace, and the technical requests your browser makes to load pages. The application currently handles these categories:
| Information | What happens to it |
|---|---|
| Landing-page draft | Text stays in the text field until you submit it. On submission, the prompt is included in the workspace URL along with the selected mode. Connecting GitHub temporarily saves the draft in this tab so it can be restored after authorization. |
| Chats and projects | Chat and project identifiers, names, prompts, source files and revisions are stored in the API database under your account. Each chat retains up to 40 recent messages within a size limit. Chats can stand alone or belong to a project; project chats share its source files. Lists are paginated. |
| Reference attachments | Selected file contents, names, types and sizes are read and saved in tab session storage for local reference previews. The feature accepts images, supported source/text files and PDFs with size limits. Selected text/source references are included in an AI task when you submit it. Image/PDF files are not accepted by the current AI task flow. References have a one-hour expiry checked when loaded, and can be removed individually; signing out clears account references. |
| Sign-in and sign-up fields | Email and password are submitted to the API for registration or sign-in. Passwords are hashed by the API. Verification and reset emails use the configured email service. Passwords are not written to browser storage. |
| Technical requests | The server receives requests needed to deliver the site, including the requested path and query parameters. Ordinary connection data can include an IP address and browser headers. Production logging and retention have not been finalised. |
| Exports and copied links | An export creates a file on your device. A copy action writes the selected code or link to your clipboard. Your device, browser, and any recipient control their own copies. |
Your browser may separately offer autofill, history synchronisation, password-manager, or download features. Those features are controlled by your browser or device provider, not by this application.
03Why information is used
Inputs provide the requested account, saved workspace and development workflow. The API handles authentication, project storage, security, transactional email, billing and task accounting. When you submit an AI task, configured model providers receive its request, source context, selected text references, and relevant prior task outputs. Authorized sandbox tests send project source and commands to Modal.
For processing necessary to provide the account and saved workspace you request, the intended UK GDPR basis is contract. Security and abuse prevention rely on legitimate interests, subject to a documented necessity and balancing assessment. Statutory privacy requests are handled under legal obligations. These choices and the actual deployment must be reviewed before public launch.
You can view the landing examples without an account. Access to the saved workspace requires sign-in. Browser storage used for requested account features is described in section 5.
The application does not implement targeted advertising, its own model-training pipeline, or selling personal information. Configured billing supports paid account management. We will explain any new purpose and establish the appropriate basis before starting it. Accepting terms is not blanket consent to unrelated processing.
04Prompts in URLs and shared information
When you start a build from the landing page, the workspace address includes your prompt and selected options. Query parameters are part of the requested address; they can be received by the site server, retained in browser history, copied with the address, or captured in infrastructure logs where logging is enabled.
For this reason, a landing prompt is not guaranteed to stay solely on your device. Do not use it for passwords, tokens, private customer data, or confidential business information. Avoid sharing repository or reference information you are not authorised to disclose.
The workspace’s “Share” action creates a private link to your saved work that requires your signed-in account. Copying the full address from your browser may include the original prompt. Anyone receiving a file or link can keep or forward it.
05Cookies and browser storage
The application does not set advertising or analytics cookies. It uses HttpOnly session cookies for authentication and short-lived cookies to bind provider sign-in to your browser. In production these cookies are Secure. The default session lasts seven days (deployment configuration permits one to 30 days); OAuth binding cookies last ten minutes by default. Sign-out revokes the corresponding server session.
| Storage | Purpose and duration |
|---|---|
| forma-demo-projects-v1 | Legacy browser-only demo records. The connected workspace does not read, import, overwrite or delete these records. You can remove them using browser site-data controls. |
| Session and action records | Expired sessions and action tokens are removed by scheduled maintenance. Verification links expire after 24 hours, password-reset links after 30 minutes. Email payloads are erased after successful sending; maintenance removes successful email metadata after seven days and all outbox records after 30 days. |
| Temporary page state | Form entries and interface state remain in page memory. The ocacity:return-to session-storage entry keeps a same-site workspace destination through sign-in and is removed on return or when the tab session ends. |
Project edits are saved automatically through the API when account services are available. A failed or pending save remains indicated in the workspace. The signed-in interface uses a session-bound CSRF value held in memory for changes. Rejecting the required session cookie prevents account access.
Use Chats to delete a conversation, or the project list to delete a project. Deleting a chat preserves its linked project. Deleting a project keeps its chats as standalone conversations with a copy of the latest shared files; delete those chats separately if you want to remove them. Clearing browser site data only removes cookies and browser records; it does not delete API records, backups, exports, browser history or recipients’ copies.
06AI providers, integrations, and training
Configured AI tasks send requests, project source, selected text/GitHub references, and relevant task context to the providers assigned by the operator. Multiple providers may process the same task for comparison, planning, implementation, or review. Model roles are controlled centrally. Quick, Standard, Deep and Max adjust reasoning and execution limits within the operator’s configuration; Quick omits optional planning and review passes. Ocacity has no model-training pipeline of its own.
Enabled Google, Apple, Microsoft and GitHub buttons redirect to the chosen provider for sign-in. The API validates the provider identity and stores its issuer and subject identifier, a display name, and a verified email where supplied. Microsoft sign-in does not assume its email claim proves contact ownership. GitHub sign-in uses the numeric GitHub user ID and a verified email obtained from GitHub; signing in does not itself connect repositories. Sign-in provider access and refresh tokens are not retained. Passwords are entered with the provider rather than shared with Ocacity in these flows.
When configured and explicitly connected, GitHub receives authorization requests and repository operations. The API stores your GitHub account identifier, login and encrypted access/refresh tokens, lists installations and repositories, and fetches the files you choose. A reviewed change sends the chosen path, text and title to GitHub to create a new branch and draft pull request. No automatic merging is implemented. Separately authorized AI tasks can execute project code in temporary Modal sandboxes. You can disconnect in Ocacity and revoke access in GitHub settings. Files you select as AI references or include in a submitted project can be sent to its configured model providers.
When you use billing, Paddle receives the email address and identifiers needed to create a customer and process the checkout, together with billing details you provide directly to Paddle. Ocacity stores customer, subscription and transaction references, payment status, and credit accounting records. These records include grants, expiry dates, task identifiers, safe task labels, project references, reservations, consumed credits and timestamps. Credit records do not contain payment-card numbers, prompts or source code. Live usage updates require an authenticated session and expose only your account’s data.
Task records store the request, source snapshots, generated results, status, model-usage records, and bounded command output so work can resume or be reviewed. Administrative activity records identify the actor, action and time; provider API keys are encrypted and are not returned to browsers. Execution records can contain code or personal information included in a task. They are separate from the credit ledger.
Before connecting live AI or other providers, we must identify what each receives, why it is needed, where it is processed, retention and training settings, and any available choices. This notice makes no guarantee about the practices of a future provider.
08How long information remains
- Unsaved inputs: held temporarily in the page. Leaving or reloading normally clears application state, although a browser may restore form entries independently.
- GitHub connections: encrypted credentials remain until disconnect or a processed authorization-revocation event. User tokens expire and refresh where permitted. Expired connection attempts are removed by maintenance. Change-request receipts contain repository, branch and pull-request identifiers, not submitted file contents, and are removed by scheduled maintenance after 90 days. Removing the connection does not delete commits or pull requests already created on GitHub.
- Account, chat and project records: retained in the API database until deletion; no automatic account-retention schedule is configured. Each chat keeps up to 40 recent messages, within a 40 KB state limit. Older history can be displaced. Deployments must define retention and backup rules.
- URL history, clipboard content, and downloads: retained under your browser and device settings until you remove them. The application cannot remotely erase copies you have shared.
- Server request logs: the application does not define a production log-retention schedule. The final hosting configuration and deletion periods must be documented before publication.
- Privacy or legal correspondence, if received: retain only as needed to handle the request and any applicable legal obligations or dispute. The appropriate period depends on the nature of the request, whether it is resolved, and whether a legal record is still needed; information should not be retained merely because storage is available.
Deleting browser data does not delete your API account. Project deletion is available in the workspace; account deletion and data-rights requests should be sent to the operator. Backup deletion periods and account-retention schedules remain deployment decisions that must be documented before public launch.
09Security and your device
The API checks chat and project ownership and authenticates requests with revocable sessions. Credentials use password hashing; session tokens and email action tokens are stored as hashes. This does not protect against every risk, compromised devices or browser extensions. Use your own account and sign out on shared devices.
Keep your browser and device updated and avoid entering confidential or sensitive material in projects. Email accounts support verification and password recovery when email delivery is configured. Backups and operational recovery arrangements have not been established by this local implementation.
The current preview is not a secure repository for sensitive data. No system is completely secure, and this notice does not give a security certification or service-level guarantee.
10Your privacy rights
Where UK GDPR or another applicable data-protection law gives you rights, those rights continue to apply. Depending on the processing and its legal basis, you may request access, correction, deletion, restriction, or a portable copy of personal information.
If processing is based on consent, you can withdraw that consent without affecting the lawfulness of processing before withdrawal. Not every right applies in every situation; any refusal or limitation should be explained under the applicable law.
Write to the registered office in section 1, marked “Ocacity — Data Protection”, and include a way for us to reply. We may need proportionate information to verify a request. Do not send passwords or unnecessary identity documents. We will respond within the applicable legal deadline and explain any permitted extension.
For data stored only in your browser, we may not have a server copy or the ability to identify or remove it remotely. You can use the browser controls described above. That limitation does not remove duties relating to information the operator actually holds.
You may complain to the UK Information Commissioner’s Office through ico.org.uk/make-a-complaint, or to the relevant supervisory authority where you live. You do not have to exhaust a complaint with us first.
11Children and automated decisions
Ocacity’s preview is intended for adults aged 18 or over. It does not perform age verification. Please do not enter children’s personal information into demonstration prompts or forms. Any service later offered to children would need a separate assessment and appropriate safeguards.
AI features generate project files and can run authorized development tests. It does not use your information to make solely automated decisions that produce legal or similarly significant effects on you.
12Updates and questions
The version and preparation date appear at the top of this document. Before introducing analytics, payments, AI providers, additional processors, or a new purpose for information, we must update this notice to describe the actual processing and provide any notice or choice the law requires.
Changes will not retrospectively authorise uses that were not lawfully permitted. We will make the applicable version available and explain material changes in a way suited to the service.
Privacy questions and requests should be directed to AURTIZ LTD. You can write to the postal contact in section 1. Please describe the information or processing involved and the action you are requesting.