open source/ free

Governance for your agent. One command.

MIND Framework adds structure, memory, and review to your AI coding workflow. No account needed. No vendor lock-in. Just better agents.

Quick start

$ curl -fsSL https://mindwai.com/install.sh | bash

Or clone manually:

$ git clone https://github.com/mindwAI/mind-framework.git && cd mind-framework && ./setup.sh

What you get

.mind/
├── config.yaml
├── agents/          5 personas
├── commands/        8 pipeline commands
├── rules/           3 governance rules
└── templates/       4 document templates

Plus shims for your editor: AGENTS.md, CLAUDE.md, and .cursorrules are generated automatically so your AI tools respect the pipeline from day one.

See it in action

We built a complete todo-app feature using the framework. Browse the artifacts to see what a real pipeline run looks like.

prd.mdProduct requirementstechspec.mdTechnical designtasks/task-001.mdImplementation step 1tasks/task-002.mdImplementation step 2tasks/task-003.mdImplementation step 3reviews/review-001.mdCode review — found 2 real bugsqa/qa-001.mdQA validation — all pass, bugs fixed
Browse hello-world on GitHub →

Framework vs hosted

Use caseFramework (local, free)mindwAI (hosted)
Solo developerFull pipeline via markdownFull pipeline + more
Code reviewYou set up the reviewer agentBuilt-in review without config
Project memoryGit historyRemembers across sessions
Multiple featuresManual coordinationCoordinates the work for you
Team collaborationGit-based workflowBuilt for teams
Setup1 command, zero dependenciesAccount + API key
CostFree, open sourceFree tier available

Documentation

Customization Guidedocs/customization.mdWriting Personasdocs/writing-personas.mdWhen to Consider mindwAIdocs/migration-to-mindwai.md

When to consider mindwAI hosted

  • Your team outgrows git-based coordination
  • You want review and QA without configuring agents
  • Project memory should persist across sessions
  • You need governance that scales with the team
Join the waitlist →