Description
SecureMind offers a privacy-aware AI workspace that empowers users to manage their AI models, knowledge bases, memory, operational costs, and sensitive data risks. This open-source tool provides significant flexibility, allowing deployment on a user's own computer via Docker or within their AWS account. This self-hosting capability ensures that users maintain full control over their data and AI operations, addressing privacy concerns by keeping sensitive information within their controlled environments.
Read More
The platform is designed to be highly adaptable, supporting integration with various AI model providers such as Anthropic, OpenAI, Gemini, Groq, Grok (xAI), and Amazon Bedrock. Users can also leverage local models through Ollama, providing options for different performance and cost requirements. SecureMind is free to use under the Apache 2.0 license, with users paying their chosen model providers directly for usage, and AWS for infrastructure if hosted in the cloud. This model emphasizes transparency in costs and avoids vendor lock-in.
Key capabilities include sensitive data detection and redaction, ensuring that private information is not inadvertently sent to AI models. Chat messages are checked in the browser, and attached files are processed to replace sensitive values before storage or submission. The system also features robust knowledge management, allowing users to add various file types like plain text, Markdown, code, JSON, CSV, PDFs, Word documents, Excel spreadsheets, and images. Images are processed by vision-capable models to extract text, enhancing the knowledge base.
For teams, SecureMind supports multiple users with distinct roles: Owner, Admins, and Members. Admins can manage users, models, API keys, and set monthly cost caps, while Members can chat and use the knowledge base. The platform provides detailed cost estimates for each AI reply and a comprehensive cost view, helping teams monitor and control spending. Deployment options include a quick Docker setup for local use and an AWS deployment for broader accessibility and team collaboration, with planned desktop apps and a hosted service for future availability.
- CategoryAI Knowledge Management
- Pricing ModelFree and Open Source
- DeploymentSelf-Hosted (Docker, AWS)
- PlatformsWeb
SecureMind's Core Features
Privacy-aware AI workspace
Sensitive data detection and redaction
Self-hosting on Docker or AWS
Open-source under Apache 2.0 license
Support for multiple AI model providers (Anthropic, OpenAI, Gemini, Groq, Grok, Amazon Bedrock, Ollama)
Knowledge management with diverse file type support (text, Markdown, code, JSON, CSV, PDF, DOCX, XLSX, PNG, JPEG, WebP, GIF)
Cost management and estimation per reply
Team collaboration with Owner, Admin, and Member roles
Monthly cost caps per user
Encrypted API key storage
Model selection based on message tiering
Optional two-step sign-in for AWS installs
Adaptable screen layouts for various device widths
GitHub Actions integration for AWS deployment
Local models for Docker installs
How to use SecureMind?
Get the Code: Clone the SecureMind repository from GitHub to your local machine.
Start SecureMind (Docker): Use `docker compose up --build` to build the app, database, sign-in, and local models.
Open SecureMind: Access the application in your browser at `http://localhost:3000`.
Create Owner Account: Enter an email, password, and the one-time setup code (found in Docker logs or Terraform output for AWS) to create the first Owner account.
Connect a Model: Sign in, go to Settings → Models, and paste an API key from a supported provider (Groq, Anthropic, OpenAI, Gemini, Grok) or connect Amazon Bedrock.
Run First Build (AWS): For AWS deployments, start the first build via AWS Amplify console or CLI after configuring resources with Terraform.
Sign In and Chat: Once a model is connected, send your first chat message to begin using SecureMind.
SecureMind's Use Cases
- Secure AI Chat
- Knowledge Base Creation
- Team Collaboration
- Cost Monitoring & Control
- Private AI Deployment
- Sensitive Data Handling
- AI Model Management
- Document Analysis









