September 29, 2026·Faq·Minds Team

Minds Platform Capabilities: FAQ

Which AI models power Minds, which tools it integrates with, how to add your own research, export formats, collaboration and sharing, image and video limits, and how usage is counted.

Practical answers to the questions teams ask when they evaluate Minds for day-to-day research: which models run underneath, how Minds fits into existing tools, what you can upload, export and share, and how usage is counted. For the broader picture, start with the general FAQ and the feature overview.

Models and usage

Minds work across providers. The AI models guide lists the current models and explains the team-level provider preference and the per-request override in the API. Teams that run their own model endpoints can connect them through team model connections.

Whatever model answers, usage is counted in synthetic responses. The pricing FAQ explains how responses map to plans, and the pricing page lists the live plans.

Bringing in your own data

Your own studies, reports and notes make Minds more specific to your market. Add them to an individual Mind's knowledge base as described in the Minds guide, or use them to ground a whole Audience as described in the Audiences guide. The Audiences guide also explains the validation score, which compares an Audience's answers against published answer shares.

Studies, stimuli and collaboration

A study can test images, videos, copy, decks and concepts across as many question rounds as you need, then export results for stakeholders or analysis. Collaborators work in the same study, and a read-only share link brings in clients without giving them edit access.

Integrations

Minds connects to AI assistants, product and collaboration tools and automation platforms. See the full integrations list or use the MCP server to run Minds from your own AI assistant.

Synthetic research results are directional and depend on the context you provide. Use them to screen and sharpen ideas quickly, and keep human validation for the highest-stakes decisions.

Frequently asked questions

Which AI models do Minds run on?

Minds run on leading models from OpenAI (GPT), Anthropic (Claude) and Google (Gemini). A team admin can set the preferred provider for the whole team, and the API lets you choose a model and provider per request. The underlying model is only one part of the result; the grounding data, context and reasoning behind each Mind drive how closely its answers match real audiences.

Can we use different models for different queries?

Yes. The team admin sets the default provider in the team settings, and each API request can override the model and provider. You do not have to manage token budgets for this, because usage is counted in synthetic responses, not tokens.

Which platforms, apps and AI assistants does Minds integrate with?

Minds connects to AI assistants such as ChatGPT, Claude, Gemini and Copilot, to work tools such as Slack, Notion, Jira, Figma, HubSpot and Google Drive, and to automation platforms such as Zapier, Make and n8n, plus its own API and MCP server. The integrations guide lists every connection.

Can we add extracts from our own studies to a Mind or an Audience?

Yes. You can add your own research, study extracts and proprietary context to a Mind's knowledge base, where it can be cited in later answers, or use it to ground an entire Audience when you create it. This is the main way to tailor Minds to your market.

Which export formats are available?

Reports and summaries export as Markdown, PDF, branded Word (DOCX) and 16:9 PowerPoint (PPTX), and structured results as JSON. Raw data exports as CSV, XLS and SPSS (SAV), so you can move from analysis to stakeholder-ready output or your own statistics tools.

Can multiple people collaborate on the same study?

Yes. You can invite collaborators to a study by email with a member or admin role. Invited collaborators can review the study and ask their own follow-up questions of the Audience.

Can we share a study with an external party?

Yes. A share link gives clients or other stakeholders read-only access to the study without edit rights. The link also shows the Audiences and Minds attached to that study, and you can unpublish it at any time.

How many images and videos can we add to a study?

Each question can carry up to 20 files, and each question round, the draft plus its confirmed run, can carry up to 100 files, each up to 100 MB. Images and videos count toward the same limits. You can run as many rounds as you need in a study.

How many images can we compare side by side?

Up to 12 variants in a single preference question, for example 12 packaging designs or ad visuals compared directly against each other.

Minds is part of