Skip to content

Data Use Policy

Version 2026-11-19 · Last updated Nov 19, 2026

Version 2026-11-19 · Mary Stella Labs, LLC

Draft. This document is a plain-language draft. Have counsel review it before launch.

This policy explains exactly how Grant Autopilot uses the documents you upload and the interviews you give.

What your data is used for

Your documents and interview answers are used only to:

  1. Build your organization's fact base. We extract text, split it into passages, index it for search within your organization, and propose atomic facts with a quote from the source. Nothing is treated as confirmed until a person in your organization confirms it.
  2. Prepare your organization's own applications. Drafts may only state facts from your fact base. Each factual statement cites the facts it relies on, and a separate check flags anything the cited facts do not support.
  3. Match and score opportunities for your organization.

What your data is never used for

  • Never shared across organizations. Your documents, facts, interviews and drafts are visible only to members of your organization. Search indexes and files are separated per organization and every request is checked against your membership.
  • Never used to train AI models — not ours and not our providers'. We use Anthropic's commercial API, under which inputs and outputs are not used for model training.
  • Never sold or used for advertising.
  • Never used as examples for other customers. Past applications you upload help style your own drafts only.

Restricted and client data

You can mark facts as restricted (for example, information about individual clients or participants). Restricted facts are excluded from drafting: they are never sent to the AI model when writing applications. Please avoid uploading personal information about the people you serve unless you need to and have the right to do so.

Optional features need your consent

  • Website crawling of your own website happens only after an owner or admin turns it on.
  • Voice: audio is transcribed; the recording is stored only if your organization turns on "store voice recordings".
  • Autopilot submission is off until an owner or admin turns it on and sets limits.

Turning a feature off stops future use. Each consent choice is recorded with the policy version, time and a hashed IP address.

Who can see your data

  • Members of your organization, according to the roles your owners and admins assign.
  • Our staff see your data only to provide support when you invite us (for example, by asking us to look at a specific problem), or when needed to keep the Service secure or comply with the law. Such access is limited to what is needed.

Retention and deletion

  • You can delete individual documents at any time; their extracted passages, search entries and derived facts are deleted with them.
  • You can set a retention period for original files. After processing and that period, originals are deleted and the extracted text is kept.
  • Deleting your organization is real deletion. We delete its files from storage, its entries from the search index, and all of its database records, including documents, facts, interviews, applications, submissions and the audit log. We keep only a minimal record that a deletion occurred (organization id, time and counts), plus billing records the law or our payment provider requires. Provider backups and logs roll off on their normal schedules.
  • Before deleting, owners and admins can download a complete export (data as JSON plus your original files).

Contact

Mary Stella Labs, LLC, New Mexico, USA · [contact email to be confirmed before launch]