Skip to content
Help Center
TopicsAPI, MCP and AI agents

Does Mindstamp use my data to train AI models?

Updated · 1 min read

Short answer

Mindstamp does not train a proprietary AI model on customer content. Model-improvement sharing is disabled for the Mindstamp AssemblyAI account. OpenAI sharing of feedback, evaluations and fine-tuning data, and API inputs and outputs is disabled. These current settings do not establish historical processing or zero retention.

The organisation AI setting governs Mindstamp MCP. Customers can revoke connected agent access. This switch does not turn off every separate AI feature.

Eligible uploads, including trial uploads, can start AssemblyAI transcription automatically. Transcripts are stored in Mindstamp. The current AssemblyAI account TTL is 30 days; different artifacts and metadata can have different retention rules.

  • Genie uses viewer questions, selected video context and creator notes to generate an answer.
  • Ask ChatGPT supplies a general answer that is not limited to the video transcript.
  • Question generation, translation, dubbing and generated speech send the content needed for their task.
  • Customer-connected external agents can retain data under their own terms.

OpenAI API call logging is disabled and audit logging is enabled for the Mindstamp account. Provider retention still varies by service and endpoint. A training opt-out does not mean no data is retained.

Deleting a Mindstamp record and removing a provider copy are separate operations. Ask [email protected] about the handling that applies to your processing profile.