Security and privacy for document-heavy work
MindParse is built for teams that want useful document AI without losing control of their files, providers, or workspace access.
What this page covers
Data ownership
Your uploaded files belong to you, and MindParse does not train public models on them.
Provider control
Use supported providers with your own keys or work locally with Ollama.
Workspace controls
Team workspaces support roles, invites, and audit visibility for shared access.
Data and document handling
Your files stay your files
- •Uploaded documents belong to your workspace.
- •You can delete documents or remove workspaces from the product.
- •MindParse does not use your documents to train public AI models.
Provider usage is tied to your workflow
When you choose an external model provider, requests are sent only as needed to deliver the features you use. If you want a more private setup, MindParse also supports local workflows through Ollama.
AI providers and API keys
Bring your own provider
- •MindParse supports bring-your-own-key workflows for supported providers.
- •Supported options include OpenAI, Anthropic, Google, and local Ollama setups.
- •You can switch providers depending on the workflow you need.
Local option with Ollama
- •Ollama can be used for local model workflows.
- •No data leaves your network when Ollama handles the inference.
- •Availability depends on your local environment and setup.
API key handling
- •MindParse stores provider keys so they can be used for your workspace workflows.
- •Keys are not reused across other users.
Access and workspace controls
Authentication and authorization
- •MindParse uses authenticated access to protect workspaces and files.
- •Workspace-level authorization controls who can access document content.
- •Role-based permissions are available for collaborative team workspaces.
Team workflow controls
- •Team plan workspaces support member invites.
- •Roles include owner, admin, and member.
- •Audit logs are available for teams that need visibility into workspace activity.
Who benefits most
Legal teams handling sensitive matters, researchers working with proprietary data, and client-facing teams that need clear workspace-level boundaries between accounts.
Security foundations and roadmap
Encryption in transit
- •Traffic to MindParse is sent over HTTPS/TLS.
- •This protects data moving between users and the service.
Storage and roadmap notes
- •MindParse stores the data needed to operate document workspaces and search.
- •Encryption at rest for sensitive workspaces is planned rather than generally available today.
- •Teams with stricter requirements should contact us to discuss fit before rollout.
FAQs
Does MindParse train on my documents?
No. MindParse does not use your documents to train public AI models.
Can I use my own AI provider?
Yes. MindParse supports bring-your-own-key workflows for supported providers and also supports Ollama for local model usage.
Does MindParse support team access controls?
Yes. Team workspaces support roles, invites, and audit logs for collaborative document workflows.
Need a security conversation before rollout?
If you are evaluating MindParse for a sensitive workflow, reach out and we can discuss your requirements directly.
Report an issue
Security vulnerabilities: Email security@mindparse.ai if you discover a security issue.
Privacy concerns: For privacy questions or data requests, contact privacy@mindparse.ai.