Startupus Semantic Core
Privacy Policy
Effective date: November 21, 2026. Operator: Startupus. Contact: info@startapus.tech.
Scope and purpose
This policy covers the Startupus Semantic Core website, private research workspace and MCP service. The current deployment is operated for its owner and explicitly authorized collaborators. It does not offer separate private accounts for members of the public. Research is used to develop keyword lists, inspect demand, group topics and prepare content plans.
Data we access and collect
- Research inputs: seed queries, uploaded keyword files and their metadata, country, language, dates, source URLs and editorial notes.
- Google Ads data, when the owner enables the approved integration: keyword ideas, estimated average monthly searches, monthly search history, close variants and advertising competition. The configured account and manager IDs are used to direct and authorize requests.
- Bright Data data, when enabled: requested queries and country/device settings, search result URLs, titles, snippets and rankings. Imported Trends files and separately requested Trending now captures may also be stored.
- Access information: provider OAuth credentials and API keys configured by the owner; private service authentication; hashed MCP access/refresh tokens; job progress and request counts.
- Technical connection information: hosting and network providers necessarily receive IP addresses and HTTP request metadata to deliver and secure the service. Their operational logs and retention are governed by their policies. Do not put credentials in URLs or research inputs.
The application does not request Gmail, Google Drive, Calendar, contacts, payment card data, customer lists or ad-conversion records for this workflow. It does not independently receive your entire ChatGPT conversation; it receives the arguments the client sends in tool calls.
How Google data and permissions are used
The Google Ads OAuth scope is https://www.googleapis.com/auth/adwords. Google describes a broad advertising-account permission; this application's implemented collector calls keyword ideas and historical metrics operations. It does not implement campaign changes. Access tokens identify the consenting Google account, and Google still enforces that account's permissions and the Cloud project's approved API access.
Google-derived data is used to display research and generate the content-planning features described here. Startupus Semantic Core's use and transfer of information received from Google APIs will adhere to the Google API Services User Data Policy, including the Limited Use requirements. Google data is not sold, used for personalized advertising, credit decisions, surveillance, or training a generalized AI/ML model.
Sharing and connected services
- Google: receives keyword requests, account identifiers and authorization credentials through its API to provide the selected metrics.
- Bright Data: receives the queries and locale/device settings selected for search-result collection. The workflow does not send Google OAuth credentials to Bright Data.
- Railway: hosts the application and persistent research storage; hosting administrators and the owner can manage the deployment.
- Your connected AI/MCP client, including ChatGPT: receives tool results that you authorize it to request. Depending on the tool, these can include saved Google-derived metrics, complete batches, raw saved captures and exports. The current connection can access all research in this single-owner workspace. The application does not send provider credentials in tool results. Configure your AI provider's data-use and retention settings before connecting; the application cannot control that provider's independent handling after you request a transfer.
- Operator access: the owner and authorized administrators can access stored research to operate the service, respond to your requests, investigate security incidents or comply with legal obligations. Other disclosures require a permitted service purpose and your consent, or a legal requirement.
We do not embed advertising trackers or third-party analytics in the application pages. Links to external sites are subject to those sites' policies. Data may be processed outside your country according to the configured hosting region and the providers you use.
Storage and safeguards
Research batches, raw evidence, source timestamps, hashes, captures, analyses and job records are stored in the service's SQLite database on its configured volume. Access to research endpoints requires service or MCP authorization. Public information and legal pages are accessible without signing in. Provider secrets are held in deployment configuration; short-lived Google access tokens are cached in server memory. Issued MCP tokens are stored as hashes. Browser-to-service traffic on the public deployment uses HTTPS; the database itself is not encrypted by the application, so volume security depends on the hosting configuration.
The workspace keeps the service token in page memory, rather than localStorage. The OAuth authorization flow uses a necessary HttpOnly, Secure, SameSite cookie with a five-minute lifetime to bind consent to the browser. No optional tracking cookies are set by application code.
Retention, deletion and withdrawal of access
Anonymous daily request/job totals are retained to enforce usage limits and contain no keyword inputs or provider responses. Research remains until the owner deletes it; automatic research expiry is not currently implemented. You may request a copy or deletion of research associated with you by contacting info@startapus.tech. Include the relevant batch/job IDs or a description, but never include passwords or API keys. The operator will verify authority over the request before deleting data; removal does not delete copies you exported or transferred to a connected client.
The authenticated workspace provides an owner-only action to delete all stored research batches, raw captures, analyses and job history. It refuses deletion while a job is active. SQLite files, hosting backups, infrastructure logs and previously exported copies may retain residual data until the owner or provider removes or expires them; the application does not promise instant removal of every backup.
To withdraw Google access, remove this application's connection at Google Account connections. The owner should also disable Google collection and remove the Google credentials from deployment settings. Revoking Google access stops future authorized Google calls; it does not automatically erase already saved research. To disconnect MCP, remove the connection from your client. The owner can rotate the service token to invalidate issued MCP credentials, or remove OAuth tokens from storage. Research deletion and access revocation are separate actions.
Your choices and policy changes
Ask the operator to access, correct, export or delete information associated with you, or to explain a transfer. Applicable rights depend on your location and circumstances. This application is not directed to children and is not intended for submission of sensitive personal information. Material changes to Google-data use will be disclosed here and require renewed consent before a new use is enabled.