SP TRADE

AI & Data Disclosure

Where artificial intelligence is used in the product, what information is sent to the model provider, and what you should verify before using the output.

Document status: transparency draft · Updated August 5, 2026 · No binding effective date has been set

What is verified and what remains open

The technical description is based on SP TRADE's active code. The operating entity, official address, registration details, privacy contact, governing law, and retention policy have not been provided, so they are not presented as facts.

1. Where AI is used

  • SP Bot Builder: Helps draft a TradingView strategy or indicator and produce Pine Script v6.
  • Blog drafts: The publishing task may create an educational article draft when an AI provider is configured; the draft is stored in the database for management.

These AI features are not connected to trade execution or fund withdrawals.

2. Information sent to the model provider

Bot Builder sends the active message set, up to 24 messages and up to 8,000 characters per message, together with system instructions. The route can use Vercel AI Gateway and the model configured for the deployment, or connect directly to Google Generative AI. The model name is returned to the interface in live mode.

Creating a blog draft sends the model provider a topic title and instructions to write educational content without live prices, news, promises, or unsupported numerical claims.

3. Storage and access

Bot Builder chat is kept in page memory and is not written to the application database in the current implementation. However, the AI provider and routing infrastructure may process or log the request under their own settings and terms. Blog drafts and publishing runs are stored in PostgreSQL and can be reviewed by system administrators.

4. Output limitations

  • AI output may be wrong, incomplete, outdated, biased, or expressed with excessive confidence.
  • Code is not considered compile-verified, backtested, or profitable without separate evidence.
  • Do not rely on the output as investment advice, a legal opinion, or a risk-management decision.
  • Verify material information and code against a primary source or an appropriate professional.

5. Choice and data minimization

You can use non-AI research tools without sending a conversation to a model. Do not enter passwords, API keys, payment information, identity documents, or confidential trading information. Starting a new conversation clears the current screen history, but does not determine what an external provider has already processed.